💡┃suggestions-web

[delicate cedar]

It's the < > option.

[autumn lichen]

suno Is there a commercial authorization letter provided?

[valid hull]

What is the best way to stop vocals when you want to extend with an instrumental solo? The chorus is repeated one too many times, so I'm extending before that happens.

[unborn condor]

I have a question. I made some music. I like the voice and rhythm so much that I'm going to make it again with slightly changed lyrics. Is it possible to re-produce it with the same feeling but by changing the lyrics? I can't find a way to do that. Every time I produce, different rhythms and voices keep coming out. Is there no way?

질문있습니다. 음악을 하나 만들었습니다. 목소리와 리듬이 너무 좋아서 가사를 살짝 바꾸어서 다시 만들려고 합니다. 같은 느낌에 가사만 바꾸어서 다시 제작이 가능한가요? 저는 그 방법을 찾지 못하고 있습니다. 제작할때마다 계속 다른 리듬과 목소리가 나옵니다. 방법이 없을까요?

[pseudo lake]

For the "Explore" feature, please add an option to copy the currently selected genre to the Create screen.

Hot damn this is exciting, is there a comprehensive list of these types of commands?

[delicate cedar]

Not this specific one. This was a suggestion by a fellow member of the Discord.

You can find all sorts of information though at https://www.suno.wiki/

The Metatags section particularly: https://www.suno.wiki/faq/metatags/

[pseudo lake]

❤️

[valid hull]

Whether a song is public or not should always be visible.

[frank swift]

search for language would be cool!!!

[feral aspen]

There is no Suno feature specifically for keeping the same song and changing the lyrics. However, there is a related feature which can sometimes do it a little bit, with some luck. In the song where you like the voice, you can try to use the "Extend From" feature to pick a time in that song after a single verse or a small section of audio and extend it from there using new lyrics (with your new lyrics in the Extended clip prompt.) So maybe time 0:15 or 0:25, something like that. You can also try using a later time but the new song may be too similar if that case. If this is not clear I can try to respond to questions using Google Translate in the Korean channel.

[remote pond]

I have been trying through at least 50 generations to get suno to follow the meter of my lyrics. When read by a human, it's pretty obvious, but Suno invariably messes it up, drawing out lines that shouldn't be. Turning 4 line stanzas into 3 line stanzas, then putting the extra line with the next stanza, thus messing up the rhyme scheme.

Is there some way we can control that, through accents on stressed syllables that belong on the downbeat, etc?

[daring sun]

When extending a clip, should lyrics from before the specific time stamp be included? Let's say current version that works well is verse one then chorus, I now want to ad a verse 2 before the chorus. So I extend from end of verse 1, the lyrics I put into the custom box, should they still have the lyrics of verse 1 as well? And is there a proper format to indicate verse 1, 2, chorus etc, so suno will know which parts should have the same melody?

[upbeat mist]

when will there be a way to download WAV's from Suno? 192k mp3's sound a bit pony 👎🏻

[tepid trench]

The sound quality is more to do with the model quality than it being a 192k MP3 file. If they gave us WAVs at this point it'd sound pretty much the same, just a bigger file. I do hope eventually when the model quality is improved that we'll get higher bitrate and eventually lossless formats though.

[unborn condor]

thank you When you first expand a song and produce it, it becomes "part 2". Can you get rid of that? Can you make 'part 2' the first song and combine it with 'part 2', which was created by expanding it again?

감사합니다. 처음 노래를 확장해서 제작을 했을때 "part2'가 됩니다. 그것을 없앨수 있나요? 'part2'를 처음곡으로 만들고 다시 확장해서 만들어진 'part2'와 합쳐서 완성할수 있을까요?

[feral aspen]

You can't do that. But that is a good idea for a feature suggestion.

However you can remove part 1 by yourself in an audio editor.

[unborn condor]

If there is no problem with the function, it would be a good idea to remove the 'part2' button of the newly created 'part2' music and make it recognized as "part1" again.

*기능에 문제가 없다면 새로 생성된 'part2' 음악의 'part2' 버튼을 제거해서 다시 "part1"으로 인식하게 하면 좋을것 같습니다. *

[unborn condor]

If I can delete or adjust the 'part', I think I can freely change the lyrics I want while maintaining the same tone and rhythm.

*'part'를 삭제하거나 조절할수 있다면 같은 음색과 리듬을 유지하면서 내가 원하는 가사를 자유롭게 바꿀수 있을것 같습니다. *

[celest bloom]

Hi, a lot of love from India! I've been using your service and it's incredible. I was wondering if I can make suggestions to make the service a lot more usable for Indian Subcontinent uscases?

[pastel lynx]

Being able to browser your song in order of pupularity is a required feature. I almost didn't notice I had 80 likes on a song.

[hushed jungle]

Could you add the option to download the music WAV files please.

Would it be possible to add the ability to download stems?

[tepid trench]

WAV files won't give any noticeable quality improvement until they improve the model quality. You'd just get a much bigger file that sounds the same.

Stems aren't possible at the moment because of how the model is trained. My understanding is that they train on complete/mixed audio which will only ever output complete/mixed audio.

I expect both of these things may be possible in time, but it's unlikely they'll be available for the current models (V2/V3)

[nocturne osprey]

Would be nice to have a seed option so I can reuse a seed I like and produce the "same" song but just edit the lyrics

[dark raft]

The explore page doesn't seem to provide any kind of information about the name or author of the track that is currently playing. If I like a song I can't find it I have no way to listen later or explore more of their work

[potent canopy]

adding musicparts in between generated musictracks would be nice, and being able to generate the same voice when you type something like [verse1 Voice a] then [verse2 voice a] and it definetivly use the same voice from the previous trackpart, that would be lovely. and adding more music instruments like a bass talharpa/ bass tagelharpa would be lovely. also churchbells or deathknells dont exist i think

[frosty venture]

Please add the ability to save voices to be reused in other songs, to maintain a personality for album-like structures.

[peak panther]

If you want to re-compose with an existing text, a very common format is ChordPro like used on ultimateguitar. I think this would be a very interesting option.

and: did anyone already mention Paypal as payment option? Or other payment methods like Amazon Pay or Google Pay?

[potent canopy]

that would be great yeah

[raven star]

Click the name of the song and it'll pull up the song page

[fading wedge]

!!! Attention!!!
A very big request. Fix the sound of hi-hat and cymbals in metal.
So far, there is often such a sound, as if on a large sheet of rusty metal poured rubble and fuck on it with a shovel. It's like a shovel hitting a huge sheet of metal with rubble on it. It hurts my head, it's impossible to listen to.

[dark raft]

Not present on mobile, but thanks

[simple spoke]

Maybe a Button to shuffle playlists would be great 🙂

[raven star]

Ah, wasn't aware! Valid request then lol

[rough zinc]

I suggest when you click "get whole song" and finished you can change tags (Style of Music). It always use style of music tags from part 1, but would be great if we can use differet from other parts... For example i used Style of Music for par 1 "sfx, sound effect", but in other parts i use "Uplifting Trance" (i needed sfx and sound effect for intro of song) and then when i finish song and "press get whole song" its only written "sfx, sound effect", but i wanna like to change to "Uplifting Trance". Maybe it's sound useless, but it would help a lot for some people (maybe just me idk, i always trying use many genres in one song tho).

[small frost]

/chirp

[rare obsidian]

The /chirp command is not available in this channel

Go to the #🐣┃chirp-beta-1 channel to use the /chirp command, and check out the #🐣┃how-to-chirp channel for step-by-step instructions for making music using Chirp on Discord.

[tepid trench]

^- Woodhouse's messages need updating to reflect that Chirp v1 is no longer available in Discord.

(Use the website at https://suno.com instead)

[vapid isle]

can we have a " start from new" in the " Reuse Prompt" option to make a Song from scratch by using the style and instruments of before ? like reuse, RE use. We cant type in 0:00 min to make it from start. when i type in 0:01 or 03 + the Song i try to make new is using the motion of the song before. hope its clear wat i try to tell. at this point i see no different between "extend song" and reuse prompt" besause both start deppending on time you type in, and they dont start from new. what Re use means to me. anyway, thanks for everything, we love you all ! osessed with suno ciao

simple "Restart song" would be epic

sorry my problem is, when i start using suno here i used songs with [verse] and [chorus] instead of [verse1] and [verse2], and some of this first songs i need to cover and make them new style that fits the technice of suno " get whole song". now i try to catch a favorite song and want to change the lyrics, so i used "reuse prompt". but when i pick a time, the song i try to create catching the motion of the song before [ the main song]. example: when the Main song is in Chorus, the " reused" song will sing his [verse2] in the motion of this [chorus]. the voice become louder and the motion is high like screaming. so " Reuse" dosent start from new what i need. sorry

and at the end of the Main Song is an "fade out" and when i pick time 1:30 for example, my [verse2] will fade out slowly...really hard ^^

[grizzled cove]

I would like too create folders in the library so i can sort my songs. Like 1 Folder for a genre or templates, for Full songs ore what ever.

[abstract garden]

REUSE THE SAME SEED
I want to re-use the seed and start a new song.

Better editing and tweaking is what I want.
I want to go back and adjust parts of the song, or change the whole beginning but keep the seed.

[tardy zephyr]

write [Outro] at the end of the lyrics.

[tardy zephyr]

I want that too, with fixed lyrics, it's a feature that needs to be implemented. Or they could just implement editing of a song?

Suno is so cool, but it's like a toy due to the RNG, you will never get what you initially want...

[obsidian marten]

I would suggest that the generation for paid members includes a midi-download feature, where a midi version of the tune created can be downloaded and imported to a DAW and turned and/or tweaked into a song by the end user.

[topaz fox]

You guys need to train some Beach Boys style harmonizations and stuff like that, I want to have Brian Wilson in a box lol

[tardy zephyr]

Also they need to implement merging different parts even tho they don't belong together by their extension method.

[heady viper]

Hey there. Love for the team! Thanks for putting this out there.

If I may suggest:

• count listens (may be happening already?) and likes (not happening) also for people that have not subscribed

This will allow a title to become popular if it was published outside of Suno but linked to the Suno page.

So it generates traffic and is a good PR/advertisement for the platform.

Thanks and best

[real lynx]

It would be nice to be able to select the gender of the voices, maybe with a toggle or something. I keep asking for male vocals but I keep getting female vocals.

[tepid trench]

This might work for the most prominent melodies (e.g. lead instrument) but would be very difficult for other parts. Percussion for example might not map 1:1 to MIDI drums, and backing instruments often don't have clearly defined notes.

I'm pretty sure they don't generate any MIDI and output audio only, the closest you can get is to use some audio to MIDI converter, but again this will only work for the most prominent instruments.

[obsidian marten]

depends how they generate the music, and I dont see why they can't. It would be a lot of fun for people that know how to do that kind of thing

[near drum]

Wish there was a way to regenerate the song art

[obsidian marten]

just take the prompt and still it in ai art generators

add "album cover" to the end and see what you get

[near drum]

that wont change it in Suno though

[obsidian marten]

ohh i see what you mean

[near drum]

If I want to share my song from Suno it'll be a picture of a weird bird that has nothing to do with the lyrics or vibe at all. makes me not want to share it

[tardy zephyr]

Suno needs to show track length, also Suno needs better library managing and sorting, it's a nightmare with 40+ pages of mostly worthless generations

But I made some really cool tracks.

[crystal frigate]

You can use playlists as "project folders", they don't have to only contain your finished, publishable things.

It's not the only possible management approach, and maybe they will add more. But it might help in the meantime.

Also downvote your useless gens to hide them

[tardy zephyr]

Of course, I am doing this since day 1 🙂

I noticed you get refunded credits if the generation is too short

But there are 1000 worthless generations that didn't stick to the original melody

where is the point to continue a track where the ai can't recognize the same lyrics and sing them in the same way as in the previous parts, with some exceptions when it recognizes shorter choruses

And I wish so much to be able to edit the lyrics to do some corrections

very minor

[tepid trench]

As I understand it, the model is trained on audio clips, and generates an audio clip in a single shot. It's very unlikely that it uses MIDI in the process. More likely it's just an algorithm that knows how to generate the waveform based on what it's "heard"

[tardy zephyr]

also as I am sorting through my library there in no way to tell if the track is in a playlist. amomg many generations with the same name I need to remember the cover picture of a track, but even they do repeat

so yeah displaying the track length would be quite useful.

and a search library function while at it

I have 40+ pages full of generations

[tardy zephyr]

I am making an Italian song and sometimes the AI goes completely crazy singing italian lyrics in spanish or generating some gibberish, I think we need an option to chose the song language

Because it totally can sing in perfect italian

[vale trench]

can we get some kind of way to force a song to end

right now the best trick ive found is to Get Whole Song before extending, it seems to end it when its gone on long enough

[distant latch]

Anyone know if there are plans to implement growls and screams for metal?

[primal yacht]

What's the trick to get him to sing as a solo, it always comes out as a group of people it's frustrating

[ember scroll]

can we get some more common payment methods like paypal

[zenith kettle]

please, for the sake of god, let the ai be able to make 5 minutes long songs, so many good songs that get randomly cut at 2 minutes, and when you try to extend it, yeah, it is good, but sometimes there is a small cut and you cant merge those two parts together. PLEASE

[tardy zephyr]

try acapella

[hybrid crane]

great advice to @primal yacht but it should be spelled a cappella, else it will not work

[tardy zephyr]

Honestly It's a nightmare

[tardy zephyr]

yes you are right a capella

[hybrid crane]

and 2 PPs

[tardy zephyr]

sure 😅 but the ai understands

How can it just cut in the middle of an incredible sung verse

[hybrid crane]

No no.. I tried many times.. lol.. 😉
With a cappella it's really keen

[tardy zephyr]

how can't it reproduce the exact same melody, if the lyrics are the same

[urban sleet]

Everything being choirs is a new V3 bug. Will be fixed soon, probably.

[tardy zephyr]

or sometimes it does but 30 other times it doesn't

[hybrid crane]

technical questions belong in #📚┃suno-school
This channel is for suggestions
Please keep that in mind 😅

[tardy zephyr]

or finish a song brilliantly, but then adds some weird instrumentation like it's another song

[primal yacht]

🙂 Noted

Thanks this is closer to what I need it is indeed a nightmare in v3 now.

[tardy zephyr]

what is the difference between v2 and 3?

[humble grail]

Hi, I wish I could choose among the covers generated for the same song.
Sometimes I only get that bird drawing because the song didn't end the way I wanted and I had to edit it the last seconds.
And sometimes I only want to choose among many great pictures. 🙂

[quasi osprey]

One thing I find is that Suno often has difficulty 'figuring out' the properly stressed syllables when converting poetry to song; it would be nice if there were a way to tell or at least nudge it towards particular stress patterns.

Even just writing the word or syllable IN CAPS or something....wait, that doesn't already work, does it?

[tardy zephyr]

I found out some voices work better than others when speaking words, for example I have a very long and complex sentence to sing, some voices do it perfectly, when others skip and cut words they are speaking

while catering to the melody...

We need a feature for the AI to remember reuse or load if you will the same melody pattern when it comes to extensions. Sometimes the AI does an excellent job, but it's quite rare.

Usually the last extension I'm making the original melody is almost gone and ends up singing the same words in a different melody and making it feel like it's a completely different song all together...

Except when it's repeating the lyrics from the previous extensions, but it does that at random.

[near drum]

Another suggestion would be the ability to edit the lyrics. Sometimes the AI doesn't complete the lyrics but when you extend the song it keeps the lyrics there that it never ran through. Also sometimes the AI adds its own lyrics. Would be nice to clean them up

[vale trench]

suggestion: better view for Extend where we see basically a tree splitting from a starting song

it would be a much better way to work with extending a musico instead of seeing them all in a list

[dawn hinge]

FIRST IMPRESSION : I had the opportunity to test version 3 Alpha and version 3 and I preferred version 3 Alpha for metal, math rock and accidentally created tracks.
I don't know if I'm the only one in this case or maybe my prompts are bad but I used all my credits on the new version and it didn't convince me for these styles of music.
However, on version 3 Alpha I got some really good results.

In terms of suggestions :
I think it would be cool to make a "create a similar song" option.

Also the possibility of generating choirs in instrumentals (for example Bulgarian or opera choirs but without lyrics) however I found an alternative by putting "aaaah" or "Ooooh" in the lyrics. The program doesn't understand what a Bulgarian choir is, though. Having the possibility to generate samples with choirs would be a plus, I think.

The possibility of having separate tracks (drums, vocals, bass, etc.) and being able to download them would be a great option.

Having the BPM of each piece of music created would be a plus.

And also an AI that understands what Nu Metal means haha.

It would also be nice to have a tutorial on the prompts recognized by the AI or the music styles, so you don't have to spend all your credits on tests haha.

[rustic gazelle]

Felt like I'd add this into the suggestions after experiencing it myself and seeing it discussed in numerous different channels frequently: White Noise

As a feature request for v4, please consider eyeing into what's causing so much distinctive white noise in generations and patching it. For reference, "white noise" has been discussed frequently since v3's release ~3 weeks ago. For me personally, it's come to a point where I only really generate songs where white noise is fitting (dark/creepy vibes).

I ain't no audiophile, but it's gotten to a point where I can distinctly hear the white noise in nearly all generations that I listen to. I know I shouldn't have my expectations in the clouds, but I'd just like to drop this off as a request to look into for future versions. :)

[leaden marlin]

Would also be more confusing

[vale trench]

how so

[leaden marlin]

I mean for people with autism it would be harder to know which is which

If it's all over place

[glossy kelp]

I think the white noise effect comes from the [Chorus] effect the non custom mode generates to make a good song (probably from feedback)

I get brilliant songs when I get gpt 4 to format the song for me

[leaden marlin]

Rip

[glossy kelp]

Seems you have to use alot of [Intro] etc into custom lyrics with description of what you want...

[leaden marlin]

Partially the case

[glossy kelp]

AI inserts [chorus] mode partially but then other part of the network thinks that does not fit and mutes it to noise levels

(the one with feedback authorization)

[leaden marlin]

I studied BTS music and there lyrics are put together

Like

[Verse 1]
So, I'm going to hate,
I'm going to hate you
Paint you like the villain you never were

And white noise is cancelled out

Through the entire thing

So each text is close together

And to start a new text you press space
And type after that

[glossy kelp]

This is what you get from what prompt? 😄

You need to do custom mode

And get it formated

[leaden marlin]

When I tested it on suno
Once I just formated the lyrics customly through discord

And directly back to Suno site 👍🏼

[glossy kelp]

I think just the lyrics is not enough to control Suno AI, you need to format it [musical terms]

Need to go to custom mode and just not enter lyrics I guess

[leaden marlin]

Depends on type of music
You can Control the beats

Like for example
You type k-pop it's going to do upbeat regardless and some different basses

And for others it vary
My point is It can be a normal lyric
With [Bridge], [Intro] [Verse 1] [Verse 2], [Outro]

But for just Beats

[K-pop Mid-Beat]

[Rock-Upbeat]

Yeah 🥰

[tough oxide]

Random play button for playlists.

[leaden marlin]

Oop

[frosty venture]

I wish we could see a deconstruction of how a songs iterative elements are chosen when choosing a genre so we could more accurately select prompts to trigger an iterative output that fits the narrative design and concept we structure our lyrics around.

I just want to reign in some of the randomness of the AI's slective process.

[glossy kelp]

I wish the song description box would be as long as the allowed prompt (in desktop web view)

[rose barn]

Add a feature were you can put (bass line starts playing) or (start solo here) etc.

[woven roost]

make the ability to generate one song for 5 points, so that way refunds for one song will actually matter

[sly grotto]

I would suggest editing the lyrics after the song is ready, so you could remove the lyrics that were not used/sung from the song when you continue the song before the lyrics part.
This way you wouldn't end up polluting with unused parts. (in suno.com)

[shy field]

It would be really cool if suno took some audio that we recorded as a sung snippet and made a song out of it. Or turn a whistle into another instrument of our liking.

[errant bluff]

I really want to better understand prompt and prompt language, it will be good to have a way to set full lyrics or part of lyrics, and change the size of lyrics. Also, it will be better if possible to provide a sample (play yourself) as a base. And yeah, we need some sequencer on top of this.

[inland flame]

I've been loving Suno for the past week, but there's a few mandatory additions that I would look into:

• Having a search function would be VERY helpful. Unless you have the url to a song (or saved it to your playlist), there isn't a way to search for a specific song generation. Users can easily listen to songs days after they release.

• With the recent profile change, some kind of follow function would also be appreciated. If a user consistently likes the music someone else is generating, they can have quick access to newer and older songs. This could also work in appreciated with a search feature.

• Personalized Recommendations. This could help push music outside of the New and Trend tab. Based off the user's genre preferences (prompts maybe?)

[rigid geyser]

able to use the same vocals/ seed from one song to another

[warped orbit]

Suggestion: Adding the ability to download an .srt file containing the subtitle timings for the lyrics. I can do this through vocal separation and transcription, but I always need to do manual editing. Would be pretty cool to add something like that. Just a thought, you prolly got lots more to do.

[hollow locust]

Are you continuing from a 'completed' song? Maybe move the timer back a bit. That's what usually happens to me.

[cosmic tusk]

Paid accounts need to have a longer length available as standard...a minimum of 4:00 rather than 2:00.

[warped hearth]

Would be very nice, to reproduce the song again with a Seed-Number and the prompt. Of course should the used Seed-Number not to be public, so nobody is able to reproduce a song from the community. But to start over again to a song "I" made with Suno with different Lyrics. Sometimes I made a song that I like, but made mistakes with the Lyrics and Meta-Infos in the Lyrics. Only with "Reuse prompt" I am not able to reproduce the same song. Maybe if you offer a button like "Reuse Song" or "Start this song again" or something like that. Would be very helpful and is very appreciated.

[pearl condor]

I should have deleted this message, I figured out the issue thanks I found out

That was my mistake, I thought this channel was general for a second

[potent folio]

Allow for more customization of artist pages ie socials

And also, ideally, allow artists to pick which songs are on their "profile"

[soft geode]

In each song, add comments, so that people can either judge or suggest improvements

[torpid skiff]

#💡┃suggestions-web please implement a MIDI export and additional a track split in VOCAL, DRUM, BASELINE etc.

And maybe some other voice attitudes like, old, young, clean, dirty 🙂

[drowsy mortar]

If you extend the song from 0 seconds, does that work (didn't try it, but that's what I would probably do to get the same "seed")

[subtle barn]

Would be amazing if it would be possible to stich certain pieces from generated variants of one song together and then re-render the entire track for cleaner result.

Currently, the more times you try to extend each extended piece - the more distorted and wobbly the sound becomes.

[leaden marlin]

It would be good if there is search bar by genre on suno explore tab.

[ebon vault]

Update song titles on videos, or don't begin making videos until the user first requests it, cos, a lot of songs come out different than expected and a name change is great, but then the video name isn't matching,

[glad skiff]

I'm missing the possibility of generating Tracks without specific Instruments/Elements f.i. a drumless backingtrack to jam on it in the rhearsal room. I tried different promts but it always generating drums along with the track even though I typed in "drumless" or make this sing without drums.

[quasi merlin]

I might be using it wrongly, but in the Beta Explore, I find a sound I like, but there doesn't seem to be a way to recreate the sound. For example - Russian Afrikaner Folk (don't ask), I tried two methods to get that sound, I tried typing it into the Song Description - it delivered "Blues Accoustic Folk", then I tried (after a lot of searching round the wheel) directly from "What will you create" and it gave me "Electronic Pop". Neither sounded like the example.
A suggestion would also be to make the wheel searchable. I love the randomness, but trying to find something afterward is impossible.
Thanks.

[vast stream]

How about _NOT_putting autotune on the vocals unless the prompter requests it?
Exception would genres that demand it, I suppose, but it is freaking everywhere now and sounds terrible.

[true umbra]

As I've been using Suno I find that with certain Styles of Music, certain tags work while in other Styles of Music, they don't.

I've concluded that the Style of Music influences how songs are created and what and how tags can or can't be used.

Can you please confirm this?

If it is true, then would it be possible to allow users to view tag options related to that style as you create other Categories/Styles of Music?

[vast stream]

Same problem. Beatless and drumless are not recognised. They should be. Allowing negative prompting would be the other fix.

[exotic jacinth]

The one suggestion I have above anything else is probably the ability to resynthesize the song by uploading the final full version and have the AI 'listen' to the song and reproduce it in higher quality or with a more consistent flow for example

[warm nimbus]

Maybe its just me, but I could not figure out how the Beta Explorer is designed/intended to be used. Wish it is made a bit more intuitive.

[ebon vault]

Another suggestion. When I am making a song, whether it is about football using Kick often will "make this prompt for review" or if I am making a song that is about something that really should not be censored. I get it you don't want some weird porn lyrics, but the limitations, especially for paid customers, of saying things like Kick, blood, head, or even USING the (gunshot) option is blocked. I understand not wanting it to write lyrics for a user about making bombs or how to create designer drugs at home, but if we write our own lyrics and it isn't a 2 Live Crew song, seriously, let up... it is ridiculous.

[fading dew]

Would be nice to make dits to the songs like you could go back and pay cridets to tweek the lyrics and re change how the ai says them

[proud yoke]

我想问一下API调用有没有渠道啊

[waxen creek]

I have found that useful. Or extend from a few seconds in, before anything that your do not want, starts to play.

[jolly trail]

Can suno provide simple audio separation?

[waxen creek]

Generating videos for every song or snippet m ust use up valuable Ai compute power. It would help if this was channelled into the music making.

[crystal frigate]

Not currently. I've been using voice.ai site, and there are other options. Depends what you want to do.

Would be great if Suno generated with stems available, and I think such a common request that it's part of the "Suno Bingo" card that gets posted routinely. In theory the generator has additional information during generation that could make stems higher quality, at least for vocals vs instrumental splitting. However, I don't think the generator composes the music with much understanding of the instruments used, so may be a lot weaker at getting separate guitar/bass/drums/keyboard etc.

[dawn oracle]

Instead of only extending, allow it to improve the beggining. I have been having terrible begginings

[quaint wagon]

Edit song function please! - Allow you to alter the lyrics at anypoint and extend using the same instrumental.

[robust arch]

#💡 Please enable dry separate stems from songs to be available for users to mix-down in a DAW. 🙏🏼

[hollow locust]

This, but for "Add to playlist"

[muted burrow]

they should add a history system where you can see all the song you have played before

[simple slate]

Good morning,
I have a suggestion for improvement. Currently it is possible to make an Instrumental track alone, or an Instrumental track with vocals. Sometimes the Instrumental track is perfect, but there is a vocal error, extending by starting before the error does not create the same identical Instrumental track. Would it be possible in the future to dissociate the Instrumental track from the vocal track, to be able to download either the whole, or each separately, to correct the vocal track, cut, fade or other...?
Thank you.

[green eagle]

What I find in most songs is that the music far overpowers the vocals. The only way to fix this is to put the song through a mixer and separate the tracks which is very hard to do as the vocals are not clear. I would like to see a way to adjust each part of the track to the depth we want to hear.

[ebon vault]

: Suggestion :

Create an "end" mask for whole song (or single clips) where it will fade to 0 from the time stamp to the end of the clip

[rare grotto]

My suggestion is to let the ai create a whole new different instrumental by showing one from a created song.

[plush hemlock]

It's cumbersome because the voice is different every time an AI song is created. Where should I make a suggestion?

I think it would be more fun if an AI song was created with a recorded voice.

[winter crag]

Any chance we'll be able to remove the vocals so we can create karaoke versions of our songs?

[hybrid crane]

Third party software is the current solution. Like https://ultimatevocalremover.com/

[winter crag]

@hybrid crane it should be a simple process for Suno to be able to accomplish this and shouldn't need any 3rd party software.

[hybrid crane]

They are still developing..

[scenic remnant]

Anyone ever thought it’d be good to have the stems available to export from the generated track in order to import them into our favourite DAW?

[hybrid crane]

I am only providing a solution, till that time...

[winter crag]

@hybrid crane thanks! I will check it out! I also want to record my own voice using the music, that way it adds a human element to the songs.

[hybrid crane]

Many people stated this wish. But I am just a volunteer moderator here. I have no say in that. The team is aware of this wish though.. But it's a minefield. They don't want people to abuse it to record voice of people who didn't gave permission (copyright).

[winter crag]

@hybrid crane I understand that completely.

[jolly swan]

Yes, something like '-guitar', '-drums' or '-vocals' would be great!

[agile blade]

make a channel for helping newbies? don't you ever think about that?

and also choose some people to be the helper with Suno-helper role or something

man... how do y'all get a proper help up till now?

[swift parcel]

Suggestion to Suno developers.
It would be nice to have a sorting function within the playlist, e.g. sort by title, sort by date.

[agile blade]

nice suggestion there

cuz mine is a complete mess

[still parcel]

We already have this. Also should work for “add to playlist”

Use shift or ctrl for multi-select

[keen geyser]

we need discord channel for tips and tricks and getting more info about "style of music" prompt

[swift lance]

it would be nice if we could use the same clip in different parts of the full song we create

[charred flax]

#💡┃suggestions-web Any thoughts around letting paying users upload music to train their own genre, or even a public genre?

#💡┃suggestions-web And any idea if stemming is going to hit the platform?

[mossy spear]

In the bug report reasons, you should add "Wrong pronounciation" (with the possibility of specifying which words were wrongly pronounced) and "Stopped too early".

[burnt rover]

Hi, I'd suggest to add structure indicators buttons (intro, verse, pre chorus, chorus, instrumental, outro etc) that automatically add the markings to the lyrics text field for sung songs, instead of having to type them manually, due to the complexity of entering square or curly parentheses on computer or mobile keyboards

This is just an accessibility advice

Also the ability to scroll through pages in a category or the "New" section on the home page to help people discover what other people made with Suno at least in the current month, and not just in the last 24 or 48 hours or so

[drifting basin]

Add ability to do (Verse: Name) to have different voices in the song

[wraith adder]

Please add the ability to be able to chose the song time or at least a min and max time so that it does not need to be extended all the time.

[leaden marlin]

Going out on a Limb here....... Is it possible to add an option for suno to Genrate "DJ Battle Records" battle records are very simple in design. mainly consist of a sample then a 1 or 2 bar music loop with another sample in between another 1 or 2 bar music loop etc. etc. im sure you could analyze some battle records to understand.

The songs end awkwardly, for example in the middle of a sentence

[agile blade]

Extend feature exist

👍

[rich pecan]

Most of us aren't professional musicians,
So it's important to have good sound quality after generation rather than trying to edit it ourselves,
Additionally, a concern: if there's a possibility of improving the sound quality in the future, is there still a chance to enhance the quality of the already generated works?

A "normal melody with high sound quality" may be more listenable than a "beautiful melody but sounding somewhat harsh."

Whether it is the audio generated for the first time or the audio generated by "Continue From",
there may be the situation of a certain audio track becoming blurry.
If didn't notice this blurry audio track at first, then the subsequent development of the music (such as changing to high/low notes),
may cause the sound quality to become even more blurry.

However, the blurry sound quality may also bring innovative melodies.
So if can filter the sound quality, make it an optional Toggle Filter Button.

[wraith reef]

I have a suggestion... so how about you have the opportunity and also the people who use it for free to upload their own beats... and then the vocals are done on it so that you also have the opportunity to use your own beats.

[charred flax]

#💡┃suggestions-web Would be great if we could stem, because a vocal only would be nice!

[mossy nexus]

It'd be awesome if when your prompts get flagged for moderation, it would tell you exactly where the issue is so you can make the right changes.

[pale tulip]

it would be nice if other languages like finnish sound better, as it sounds sometime(almost all times) too soft voioe, like too processed. Some other maker better finnish sounding songs but melodies are poorer

[twilit frigate]

Just a short list of things, I'm sure others have brought many of these up already.

• I'd like to make one song by default, and have it use half the credits. (I cannot find an option to make only one song generation)
• Option to filter by disliked (So I can mass-delete).
• A select all feature
• Unliking a song removes it from view, but not from autoplay.
• I'd like an option "Dislike will also put into the trash"
• Pro users should be able to define the song length (To a reasonable limit)
• Deleting a song should refresh and move the list up. (If I delete a lot, the page becomes increasingly empty)

[tidal wharf]

can we make sub-channels by genres? I'd love to chat with and support fellow heavy music appreciators and check out each other's work

[quartz badge]

Can Suno please provide a comprehensive list of prompts, metatags, song structures, etc. that are known to work well? The Suno Wiki is kind of bare bones. Feels like there has not been extensive research done by the devs, or if there as it would be nice to share that with the users

[dry robin]

i am part of that genre. im suprised if you have anything to share for the last 3 weeks lol i post, i moan, im ignored, unless sorted my 30 dollars will be going elsewhere

[hybrid crane]

The wiki page (suno.wiki) is made by the community. Suno recently added https://suno.com/explore to help people understand and showcase things that can be done. Just click on any term, and then on this to see what they done:

[tidal wharf]

lots, actually. See e.g. https://suno.com/song/21fc844e-bc17-4404-9512-39fdf1190e23 (only got to filling my profile, and not everything can be put directly from suno as I do some post-editing/cutting)

i post, i moan, im ignored

what is your specific problem?

[dry robin]

love that

how do you get guitars that dont sound like a washing machine

[tidal wharf]

well, you can see the prompt in the vids, there's nothing that much special, I think

[tidal wharf]

thank you! I actually got one album released and on platforms, the second one coming soon (6 songs ready, 4 in the works): https://generated.bandcamp.com/album/divine-discontent

[tidal wharf]

I think it's important to "catch" on a good sound once you get it. I.e. if you can get a decent 5-second intro but then it goes sideways, just extend it but set 0:05 for the continuation

sometimes the best ones melodically will not be the ones with the best sound, so you'll have to trade one for another

[pastel lynx]

"guitar played through a tube-amp" Makes it better.

[tidal wharf]

also, I usually generate a continuation 10-20 times before I find something that really clicks with me

[dry robin]

yeah i have an album out now, its doing well, which is great, second album is now done but had to do last track through udio, 3rd album nearly there as well lol

[tidal wharf]

do share

[dry robin]

https://open.spotify.com/album/3PP7LIbSdRn7cjifv6e6x5?si=TL-AEbqOTvOyRA3FvHkjfA

[tidal wharf]

found it on apple music too, will give a listen today and share my feedback.

so you said you're doing great, did you do anything big promotion-wise?

[hybrid crane]

alright.. this is the #💡┃suggestions-web chanel.. Small reminder 😉

[dry robin]

appolgies

appologies

[tidal wharf]

pardon

see, here's why we need genre channels 🙂

(that is a suggestion)

#general gets scrolled away too quickly

[dry robin]

https://hypeddit.com/ its a great tool

artists still updating for today but this is as of yesterday

[tidal wharf]

wow, amazing

[dry robin]

just need to be able to produce the quality i was producing with suno

[tidal wharf]

I'll head over to #🗣️┃chill-chat

[loud cairn]

I wish i can point word stress in my lyrics. Can you make some inlyrics commands for Suno to add specific accents to arrangement?

[tidal wharf]

I found the accent marks to work pretty well (á, é, í, ó, ú). (I use Russian equivalents, but I imagine those would work for any language.

[loud cairn]

а какие эквиваленты есть на русском? а то я просто выделяю ударения капс буквой. И Суно иногда это соблюдает

[potent folio]

wait, didn't you post this earlier on another account? Or is that someone else with a similar album?

no wait

I see now, thr cover is different

it's just a bizarre coincidence

[feral aspen]

For 'stress' in the singing sense (not pronunciation) then spelllllllllling or punctuation!!! or CAPITALS or *STARS* can sometimes work

[tidal wharf]

all AI covers look the same lol! (jk jk)

[potent folio]

hmmmmmm

well in any case, giving it a listen now!

[tidal wharf]

А́ Е́ И́ О́ У́ Ы́ Э́ Ю́ Я́ а́ е́ и́ о́ у́ ы́ э́ ю́ я́

[loud cairn]

поищу как это удобнее вводить прямо с клавиатуры. Спасибо большое!

[hybrid crane]

Please communicate in English?

[potent folio]

OK, so I like the musical components better than I like most metal. However, I do think it's worth modifying 2 things, if I may?

and this is purely from a page perspective

[tidal wharf]

Its about marking stress vowels in Russian

This isn't mine, attn. @dry robin

[potent folio]

I'm putting a baby to sleep 😄

[potent folio]

removing my response lol, I just saw this, I'm lost I got pulled into this channel by a phone notification

[hybrid crane]

lol... It's not the end of the world

[tidal wharf]

I made a text note that i go to to copy and paste lol

[loud cairn]

I did the same with your message)))) Tomorrow will try it

[urban sleet]

Bring back V3 alpha.

[leaden marlin]

Hi, sorry for asking but i dont see any chirp channel, do I have to wait for adding?

[flat dagger]

Hey everyone, been trying out Suno and being a sound engineer would love to be able to get the separated tracks of each instrument to download, making me able to edit, process and mix tracks on a DAW.

[tepid trench]

Chirp is gone from Discord. The website at https://suno.com replaces it 🙂

[shadow viper]

The website is a little difficult for my friends who are blind. Especially logging in using Google / discord/ Microsoft. Does any one have any helpful tips on this? Thanks in advance!

[round elm]

I second that, well I used to be a music producer for 10-15 years, not as a profession, anyhow thats beside the point. I stopped making music many MANY years ago, but when I discovered suno-ai , Ive slept nothin almost 3 days straight, been using all of my 5000 credits u get for paying a month lol. tldr, this AI has awoken that spark, the fire inside of me, I can see myself incorporating my own music, writing and making fusion songs with me and AI, I literally have any genre, music style, I can do Gospel Opera Metal Jazz DnB if Id feel like it, or just create new instrumentals for rapping, singing, or making a song I can sample legally and actually profit out of. I too also want to be able to at least separate the vocals and the music in 2 tracks, but optimally Id like it if each instrument could be separated. I am also a bit disappointed in V3 when a song is too long, the AI adds all sorts of reverb,chorus,flanger effects as to kinda hide that the audio generated becomes "cheapened" out in the end. But theres other services that can do audio separation which Ive used in the past, Im prolly not allowed to say their apps name here. But that also costs even more money to separate tracks that when I download them are in a lossy, crappy mp3-encoded audio file. wav, flac would be preferred imo. Flac prolly would be better on their backends storage wise altho costly processing wise. When u use this app Im thinking of that separates audio tracks very well, it works best when the audio quality is top notch and these mp3's..."That Dont Impresses me Much" as Shania Twain would say hehe. Further processing on lossy audio formats degrades the audio quality even further.

But what I kinda almost demand is an option to export songs in bulk, bulk download all at once, or like google does it u can make them export the stuff, depending on how much data ur requesting it may take minutes to hours to days to get a link to zip files with all of the stuff. I cannot do like 3-4 clicks downloading like 500 of my songs, Id go nuts!

[outer vapor]

А куда делась возможность генерить музыку? Не пойму. Раньше писал /chirp и отправлял текст. Сейчас этой команды нет. Хелп )

[hybrid crane]

Чирп ушёл на пенсию, его больше нельзя использовать в Discord. Пожалуйста, используйте https://suno.com/ для создания музыки в веб-приложении.

[outer vapor]

О, нет!!!

[hybrid crane]

Но веб-приложение лучше. Дискорд-бот был версии V1. Веб-сайт представляет собой версии V2 и V3. И как бесплатный пользователь у вас есть 50 кредитов в день.

[outer vapor]

Спс. Пойду творить и созидать)

[azure chasm]

This new button yall added in the location its at is annoying. Can yall move it by the “home” button since it makes more sense for it to be there since its called “explore” thanks.

[dusty swallow]

I think in my opinion there should be an option to choose a female or male voice

Make Android app From it

Need a Android app

Official

And login With Discord in Android app

[steel fractal]

List of banned words or highlight the offending word/phrase in the prompt automatically

[onyx folio]

I wish we had some control after the music is generated, like changing it's key, changing the lyrics, vocal style etc

[pale turret]

it would be fun to control different subjects of a genre, like for example: " a hypnotic techno track with minimal house stabs/synths"

[wise pine]

I think a "to the tune of" feature would be great if someone has a parody idea - as long as the lyrics are original and don't match the real song

[leaden marlin]

Beep

Allow generation of instrumental base off an already existing generated song that have lyrics. Basically getting the stems of the generated complete song

[leaden marlin]

Adding "Key" or "Scale" or "Chord Progressions" would be a nice touch both for instrumental and lyrical, Tempo/BPM

[open vessel]

It would be nice to organize things into folders in the UI. The ability to move items from the "My Creations" list to a folder and remove them from the list would be so much better than just a big long dump of music. It's really hard to organize creations.

[severe glen]

How do I stitch/merge the extended part of the song?

[hybrid crane]

This is the channel for suggestions.. not questions. They are best asked in #📚┃suno-school or #💬┃general-chat

[severe glen]

Suggestion: Make it easier to etend the song... 😁

[proven slate]

I got so many suggestions, but main one is, can you please do something about the audio quality? Basically im making metal songs and 9 times out of 10 the audio sounds warbly. Especially the cymbals. It's quite annoying

[proven slate]

I second this. Came up with some killer lyrics the other day and said i couldnt use them due to banned words. Im like WHICH banned words

[glossy mica]

Sorry if this had been said before, but I would love a feature to save a voice from a particular track of even just the instruments, maybe rhythm or tempo too. to customize a profile around said voice and adjust. essentially I'd like to see more customizability on our end to salvage the very best pieces of a generated song, and build off of them.

[leaden marlin]

It would be nice to have a way to modify our own voices into the song
Seems as like most AI allow you to record urself and modify the voice to fit the song with the correct chords

Although me myself I only use Suno for background music but for the ones who don't it would be a good option

Agreeing onto adding the Higher Quality of the Audio and the way it skips half the song

Needs to be fixed

[slim thicket]

the banned words are too large..it's annoying.. even in French, which is the langage i use. some words can mean something wrong sometimes, but in general these sames word are very normal. ex: the word "gueule" is a word that you can use to insult someone like "ferme ta gueule" (shut your mouth) but it's the word we use to name the mouth of animals "gueule de chien, gueule de loup" dog mouth, wolf mouth.. it'S just an example , but too much words are banned for nothing.. i suggest you review your way to censor because every song i make a have flagged prompted and i must try to guest wich word or words are banned... i pay for suno app but i dont know if i will renew cause this problem is really really not fun at all..

[shadow crater]

We have soooo many Parts each song - PLEASE we need FOLDER - "create new folder", drag and drop song-snipes in - be happy that you don't be crazy because all this mixed up. And better visual view what parts are parent.

And why i cant mix version parts in the full song - maybe i want 3 parts on this parents path but the fourth from the cousin part

Than we need an option to say make here an end part of a song.

Than an option for time to wait before the singers start.

The AI is really good - but the human interface need so many more option and ways to say the AI what we want from it.

AND you will be make much money if you make a little editing tool to cut parts and put it together and than the ai make a clean crossing.
I want to say, this 10 second from this Part, this 5 Seconds from here, the voice from part 5 but with an different text.
And than we can really be all great composer. The AI are great but the AI can never feel, what the User want and feel.

Make this suno really big and great its a wonderful tool! Give us more interface please.

Have @ll a great day and i hope someone from SUNO Dev's will read the text. It is amazing what you have create - I want much more, NOW - just a joke - i give you two days LUL - sorry, just kidding. I like it really. What is realistic for what i wish - half year, year? I hope.

[north dune]

Could we shorten the song length and optionally choose to extend? I usually listen to the first 10 or 20 seconds and immediately hit the thumbs down. Every now and then I like the song almost right away.

[golden badger]

A channel for meme/funny generations would be nice

[mystic zodiac]

I gave Udio a go and found that for me Suno seems to be better at creating the styles I want more often but Udio does have a much better sound quality, especially the vocals and where they sit in the mix for stems and post-production in a DAW. If Suno were to achieve that sound quality and mix and also to add more parts to allow more control over what we get and download the stems, I think Suno would easily hold its position in the AI Music Market.

[drowsy mortar]

I felt the same way. The vocals on Udio are amazing, but the genre I like to create (synthwave) isn't modern at all (vocals sound like The Cure vs. melodies)

[ebon vault]

create an adequate file browsing ui/ux .

Create drive / explorer style folders with folders for title and arbitrary sub folders for organization before we have 500000000 files at top level we can't navigate at all

[rich pecan]

If the generated audio melody sounds good but is somewhat blurry
could set up a Re-Generate with some seed Button to create almost identical melodies with different sound qualities?

[feral aspen]

My guess is not with a seed, but that overall concept is a really nice feature suggestion

[rich pecan]

You may be right in what you're saying. I mentioned this suggestion because I often find myself confused by the following examples:

First Generation: 02:00, example:

• 00:00~00:30 # Beautiful single track, ✅
• 00:30~00:40 # Added drum loop (❌ 10 seconds of drum loop - blurry)
• 00:40~02:00 # Added bass + chorus, both perfect (except for drum loop track) ✅

I tried changing the mixed-in instrument (replacing drum loop) from 00:30 Continue From, and the new track indeed sounds clear.
However, I can no longer generate the more beautiful melody from the first generation 00:40~02:00.

[granite violet]

Where is the API？

[dawn oracle]

Suggestions:
Ability to upload a cover
Create an equalizer video automatically
Ability to delete the prompt and edit lyrics before creating video

[leaden marlin]

Filterspeak and phonetic spelling of words

[leaden marlin]

Male singer, Female singer in the genre

[indigo patio]

A repeat section would be good. I find I want to repeat the chorus of a song but the ai makes it sound different each time

Also a lyric editor wouild be good

[quick belfry]

That Would be nice that AI follows the lyrics when we enter some. i made a lot of songs and put them to trash because the lyrics that were sang by the AI were not the same that I entered.

[leaden marlin]

you have to prompt engineer them a bit.
I have also seen this show up if you are continuing a part at an inopportune time. Try moving your injection point a few seconds earlier or later

[brazen wing]

Yes I do agree, if the ai have made a good structure of the song, and you need to regenerate the lyric you are fucked.. and be able to add before and just not after also

[tepid trench]

There is no API at the moment.

[sudden pumice]

Howdy!
Would love if we could edit lyrics after song creation as sometimes the AI will alter the wording or repeat phrases. No problem with the alterations as it almost always adds to the song structure and is appreciated.
But yeah, altering lyrics after generating a song would be swell.
Thanks for the awesome tool, Suno team.

Oh, lol @the goon in jest already mentioned this. My b.

[crimson ivy]

I made this cinematic songs with UDIO and I never managed to get THIS quality with Suno. Will there be an Update regarding sound quality?
https://filmmusic.io/song/12289-the-princess-in-the-big-castle-film-score-feat-udio
https://filmmusic.io/song/12290-the-desert-planet-feat-udio
https://filmmusic.io/song/12291-epic-drone-shots-004-feat-udio
https://filmmusic.io/song/12293-violin-in-space-feat-udio

Regardless of the genre, UDIO actually beats SUNO by 500% in case of sound quality. SUNO still has many hearable artefacts and very bad quality with higher tones like high hats or strings in general. Also listen to the stereo field.

[pulsar dagger]

Day 1 asking for API.

[noble frost]

Not sure why they're so slow with the API but been asking about it since launch and havent heard anything :/

[vapid isle]

for Extensions: options like " make unique" so we can split the parts and change the hirachy. And / or Copie Extenions. I have a nice chorus part and i want this to be also at the end, but at this moment i need to spend a lot of credits to find by coincidence the same motion/ instruments to match the full song motion. simple copie paste an extension would help so much. i know programming is so much headaches, thank you all making this real here. ciao have great day

[noble frost]

Agreed - the sounds quality in Udio is way, way better. But the song compositions not so much. If we had a tool with sunos arrangements and udios sound quality that would be amazing

[crimson ivy]

Yes, true.

[pulsar dagger]

If is because of the profit, tax the API higher to ensure that your suno app is profitable, but they will 100% win more from big implementations, there is a huge market

[prisma birch]

Hello everyone,

How can I incorporate Japanese instruments into my songs exactly? Simply adding Japanese instruments with [prompt blablabla] doesn't seem to work.

Additionally, I encounter the issue that some tones/instruments are too dominant and loud in some songs.

I feel like I might need to remix all the songs with an external program.

the prompts u are using can be endless?

[hybrid crane]

Thanks 😊

[peak vale]

I think i would a game changer, if u can add steams to the tracks,

[quartz steppe]

I have several requests and suggestions. My problem is that I am aiming for quite extreme genres, that are far from mainstream. I started playing with Suno about 3 weeks ago, and during first one I couldn't get anywhere close to my promts and suspected, that model simply doesn't have the right dataset. So, eventually I got desperate and reported one track, stating the issue and asking to give model more data. Few hours later I got the first result, that fit my prompt, though later it fell off from genre for a long time. So my experience lead me to some ideas, that IMO might help with the process:

• It would be nice to have some guide about how Suno process prompt and like/dislike feedback, and how it is related to tags, that Suno puts to tracks it generates. This is needed to get an idea on how to use current mechanics properly.
• In past 2 weeks I managed to significantly increase frequency of tracks, that fits the target genre, style and mood. But, when model gets there and generates tracks based on right dataset, it hallucinates heavily, since it is not trained for this yet. At the same time, if I evaluate tracks basing on quality, there is a high risk of reversing all progress with adjusting model to genre. So that creates demand for an optional detailed feedback on tracks, where I could give distinct scores for everything specified in prompt, like genre match, vibes, tempo etc. In addition, a general quality also needs to be indicated separately. In such way it would be possible to tell the model, that track is perfect match for the prompt, but it hallucinates heavily and quality needs to be improved to be rewarded with higher score.
• It would be nice to have some channels dedicated to specific genres here in discord, where I could share results, ask for advice, cooperate in some way. Currently I feel like I am alone with what I am trying to achieve, even though I making some progress.

[leaden flame]

Feature request for Suno: On the Create and Library screens, please show the audio clip time length on the audio clip bar (row). Also the date & time created would be very helpful as well. Thanks!

[outer flare]

At least tell us which lyrics are flagging the song, it's very pg-13 level stuff, nothing too bad, and it still won't go through.

[tribal sentinel]

Edit a song you've already created with Suno. Keep the melody but change the lyrics. Keep the Chorus but change the melody of the verse. Keep the song, but move part A to the end or other way round...

[willow gulch]

any chance we can hide songs on creation page so that they will still be available in the library? I prefer to have clean Creation page when im creating a new song

[tulip beacon]

Hello! I intended to use Suno for doing small animations. 2 minutes is way too long for amateur animation. Is there a way to generate only 5 to 15 seconds?

[atomic pulsar]

hi when will there be creation of songs for 3 mins long instantly. and not like random durations.

let there be a choice of duration creation for premium members

[visual brook]

bark is probably better for your use-case, in this very limited corner case

[storm wedge]

Soundqualität lässt sehr zu wünschen übrig.
Vielles klingt sehr schwammig und verschwindet im Hall
Meistens werden Stimmen zu sehr durch den Vocoder gejagt, so das eine STimme alleine nicht mehr zu hören ist.
Das ist ja ein ganz netter Effeckt... aber so etwas sollte nicht auf Dauer eingesetzt werden!

Fazit:
Weniger Hall
Weniger Harmonizer!

Sound quality leaves a lot to be desired.
A lot of things sound very vague and disappear in the reverb
Most of the time, voices are chased through the vocoder too much, so that a voice alone can no longer be heard.
That's a really nice effect... but something like that shouldn't be used long-term!

Conclusion:
Less reverb
Less harmonizer!

[short vigil]

Mods, what's the chance that you can create roles for subscribers to have, which is auto synced via the suno website?

[crude quest]

This maybe a stupid question. is there a way to add the extended part to the 2:00 mins of part 1 in Suno?

[prisma eagle]

This you mean?

[serene sundial]

UI: When you add a track to a playlist, the notification that it's added should be in a different position where it doesn't overlay any interface elements like the timeline and playback controls.

[crude quest]

Thank you, thats it

[tulip beacon]

Thanks!

[raw garnet]

Having so much fun playing with Suno. I would be amazing to have the possibility to get stems of create tracks thats one. Two I think if there would be a possibility to upload your own voice onto Suno, it would 100% take Suno to another level. Having a possiblity to download a wav file would be nice too.

[vagrant sigil]

Can you make an option to expand the same song without changing its style and keeping the same idea behind lyrics in a seamless way?
I would prefer to make a 1 longer song, for the cost of 3-5 short ones or even 10.

[kindred hull]

I think that the MOST needed change is to be able to extend the song 2 minutes. It may be optional, (with 1x and 2x options), maybe consuming double the credits, but extending a song is too burdensome, even more when sometimes the first song generations only yields 1:20 songs, which require at least 3/4 gens to get a 3:00 minutes song.

[carmine quest]

Music quality is good, but I find it too repetitive, especially for instrumentals only generations, so it gets boring after 30s.

[shut pivot]

great idea

[fringe knoll]

No, I'm sorry but this is not correct.

AI does NOT NULLIFY human-authored copyright. That is silly. Human rights don't disappear.

If my Suno song uses my original lyrics, my original lyrics have a copyright. AI does not stop me being a human. LOL

If extensive edits and post-production is done on the song (adding/removing instruments, re-arranging, re-recording) the human-authored parts of the work are eligible for copyright.

Copyright is NOT all-or-nothing. Copyright is NOT subtractive. I do not lose my human rights just because you post some clickbait links from an unsourced blog.

Please consult the ACTUAL copyright office in your country, not some unsourced blog post.

The US Copyright Office offers guidance on how to submit works that incorporate AI
https://www.copyright.gov/events/ai-application-process/
I'm sure if you bother to look, you will find similar info in your country.

Take your hour in time out to learn from the ONLY valid information on copyright law: your government's copyright office.

[atomic pulsar]

please have a choice to generate with MALE voice or FEMALE voice. it keeps giving me female voice when i want male.

[tidal wharf]

Why do downvoted tracks no longer disappear in create? Tell me this is a bug not a feature 😭

[night bane]

You should definitly add a search bar for music categories, it's so cool to listen to all kind of music, but you can't choose the style, only the ones in the main page. Suno is a great plateform which could become very big, but you could turn it into a real music plateform, although it's generated by AI.
So add a search bar, it would be incredible

[timber totem]

What be the jazziest genres

[somber sparrow]

A feature I would like would be the ability to save a part an inject it into generation. Have the perfect chorus save clip and select mesh on that part when you want the chorus to come back etc.

[sonic flint]

feature request: Each generated song has a seed number. If you copy that seed number into a new song, you get the same melody/voice/rhythm, but can fix/adjust the lyrics. Kind of same way it works with generative images in Stable Diffusion.

feature request: More standardized tags, labels, keywords, like [intro], [bridge], [outro], [rap], [shouting] to control the singing style. And some specific behaviour for dots, commas, line breaks/endings.

[lament citrus]

I would love a feature that would allow us to choose how long we would like the generated segments to be, to give creators a greater sense of control over the "movements". This could allow for say, a dance or rap segment, in a song of a differing genre.

Bulk deleting of songs would also be very helpful to clear out the library of fragments.

[misty parcel]

Can we get the option to download an instrumental and the lyrics separately? Can we get an option to teach the AI our lyrical style?

[normal atlas]

Allow us to edit the lyrics that are attached to the song, the AI sometimes misses or repeats lines and it still sounds good, but i'd like to make the lyrics match the song itself 🙂

[misty parcel]

Can we can get an option to insert our own audio and have a song created around that?

feature request: I have cousins who are producers, and they want to incorporate Suno into their creative process, but they keep asking if there's an option to insert your own beats, lyrics, and other audio and create a song around that.

[mystic zodiac]

@misty parcel I like your idea on teaching the AI our lyrical style. Perhaps something like a keyboard that we select the melody of the vocal?

[pastel lynx]

Change the font, that's gross.

[ruby stag]

Suggest changing the rules on this server to read "no profanities" instead of "limited profanities" if people are going to be moderated for saying goddamn

[hybrid crane]

It's more about respect, and also keeping it English. But thanks for the suggestion.

[stuck steppe]

hold SHIFT and you can select multiple songs

[finite sedge]

#language-suggestions
please consider Português (Portugal) and change the existing Português to Protuguês (Brasil), there are a lot of differences

[hybrid crane]

I just created one for you: https://discord.com/channels/1069381916492562582/1230630513299947581

[finite sedge]

Thank you

[leaden marlin]

Could be possible to have 3 or 5 minute music if we choose instrumental ?

[tough bluff]

https://app.suno.ai/me: Suno

[lyric plank]

Itd be cool if we could choose the main picture for an extended song. I really liked the 1st picture suno gave me, but not the ones for the two extensions i added. Unfortunately the completed song took the most recent picture which was not my preference

Also feedback: 50 credits gets you 10 songs, 5 credits per song, right so why not drop the whole 5 credit per song thing to 1 credit. (Sorry Im not sure how to word thus correctly but...) 50 credits = 10 songs = 10 credits. Im not saying give us more free stuff (you guys are already doing amazing in that department) im just asking why not call 10 credits '10 credits' and simplify things (1 credit per song). does that make sense? 🤔

[wind crest]

i request this frequently haha. especially since for the final section of songs where often the only prompt is [fade out] or maybe a closing line or two, the image just always ends up blobby paint colours, some artsy lady, or a bird with a guitar.

[leaden marlin]

One message removed from a suspended account.

[short solar]

this agreeing with Moja i want to be able to figure out how much time i want in the extended too many times it gets cut off and i have wasted all the credits

[halcyon veldt]

I would like to use the exact same voice, instrument, theme from an other song I generated with Suno but with other lyrics

[main mango]

Same here. I’ve got multiple “voices” I’d like to express at length

[halcyon veldt]

The generation is too random. I've got the lyrics but too much variables changes. I would like to pin some down.

[grizzled meadow]

i was wondering if you could add a way where you can describe what the album/song cover should look like.

[frosty breach]

Play speed would be neat.

[latent phoenix]

How i can get 150 credits?

[hybrid crane]

You cannot get 150 in one day.

You get 50 per day

You need to use 50 to get 50 the next day

[latent phoenix]

I sighn up, but still 0

[hybrid crane]

Free daily credits refresh at approximately 10pm EST, 3am UTC.

[steel fractal]

Can't you do this by extending and just adding the lyrics you want then later editing the song by cutting the first part off?

[halcyon veldt]

Nop it doesn't work

[safe oracle]

We definitely need ways to limit or set certain times for generations. Extending songs when a beat hasn't fully finished often leads to hallucinations of the ai or it completely skipping parts or entire verses even. Sure it can be compromised by counting beats but not everyone does know how that works and if you're off beat once a lot of generations or extensions break and generate like 15 seconds of music. Its a huge waste of credits sometimes.

[odd glacier]

Use "male vocals" or "male vocalist" in your prompts.

[odd glacier]

Not sure what the convab was about but Suno's own website states that if you're a paying user, you own the rights to everything you generate - music, lyrics, etc.

So even songs with AI-generated lyrics are copyrighted by the user (as long as they're a subscriber).

Personally, I'm a bit of a romantic when it comes to AI, so I like to credit any AIs for their contributions (lyrics, vocals, instruments, etc)... just covering my bases when that robot rights movement comes along lol

But I also write my own lyrics now.

[prime spire]

please add ability to set how long you want extensions

I usually want 10sec extensions for transitions just a big waste of credits.

please add ability to set context window start time when extending.

sometimes I just want it to take a certain set of time into account

please add love button where if I double click like it marks it in a different way

this is a ripoff of 500px before they removed it and ruined the website due to VC stuff

please allow for listening to full song in order on any extensions without having to glue the whole track together. slight pauses are fine.

[leaden flame]

It would be nice if paying members will be able to specify both the original song clip length and extension lengths. The more control, the better. 🙂

[prime spire]

I figure most mainstream people on the app just would want a "make entire song in one click button" that keeps extending till it hits an ending.

behind the scenes after 3 some minutes add a "ending" tag

also another one which is probably wholly incompatible behind the scenes, if transitions were first class specify tags for start, end, time length.

interpolate the weights of the tokens linearly

[frosty breach]

If chirp isn't used anymore, could chirp-technical be used for suno itself? (If it already is, perhaps a name change?)

[stiff dome]

I dont know if this suggestions have been made but I think this would be useful:

Add a reference song to get something similar it could be a link from other suno song

Add the ability to edit/recreate in between a song, for example regenerate between 00:22 - 00:35, sometimes I get a good song but if there is a part I dont like at the beggining or pronutiation is wrong at 20 sec, then the whole 2 min song that I like is gone.

[inner rover]

It would be nice if suno made an update were you could change the album covers to your liking, and manage the lyrics description of songs because it often gets messed up while mixing songs on suno.

[latent spoke]

How do I edit lyrics or an instrument in an audio generated on Suno?

[tidal wharf]

You can't. Best you can do is click “Extend” and enter a timestamp before the moment you want to change

[latent spoke]

Thank you for responding. I still don't understand, can you explain in another way?

[finite slate]

need a search feature to go through your clips

[latent spoke]

How do I edit the audio made by Suno? I want to change the sung lyrics or an instrument within the same audio.

[finite slate]

put it in another audio program that lets you edit it

suno wont give you that kind of control

[latent spoke]

Do you know the name of any app that allows this kind of editing?

[pearl condor]

Search bar to find users and songs i think would work well

[latent spoke]

?

[fervent sable]

Enable users to create sound effects in addition to songs and music ig

[woven wing]

– Ability to not only "Extend from a time", but also redo any specific part of a track (like from 00:30 to 01:10)
– "Metatags" in theory sounds like a really powerfull feature, but it doesn't work for me in most cases.
– Genre mixing is rarely works if genres are far from each other. Let's say "Renaissannce Music, Surf Rock" generates either renaissance, either surf, not mixing it tigether.

[primal basin]

Ability to clip a portion of a track off at the end, delete an unselected portion, and extract a selected portion into a new snippet entirely.

[balmy zealot]

~Don't let us put the same song into a playlist more than once... Particularly frustrating with older no-artwork stuff that we had to skip to Page 400 to get to and the page numbers keep changing because the order is backwards ~

[primal basin]

Ability to generate only 1 version for 5 credits instead of using 10 credits every generation. THIS WOULD BE HUGE and yet probably lead to some form of lost revenue 🤷‍♂️ .
I have 5 extra credits on my account because one of my generations was too short and it refunded it, but now I can't use those 5 credits LOL.

[balmy zealot]

There is no URL to click to get to songs that have no title in the "Extended From.."

Adding to playlists should have a "recently Used" at the top.. taking 20 seconds for every song I make every time to scroll down to the 50th playlist is dumb

And dont make the playlists so huge, too many pages! I can't find anything

This is normal 100% size :

Custom reorganization of playlists.. or at least A-Z

[hallow niche]

Leaderboard in Discord and website with highest members

[solar pond]

Suggestion
Ability to hide the style / prompt when downloading in video form

[vale venture]

I'd like to see if possible generating songs without artwork for reduced credits, we know the art isn't free but I also suspect a lot of the time it maybe irrelevant if people are planning to do their own/or just don't care about it. 😈

[royal cedar]

When getting a “whole song”, allow users to specify which cover art will be displayed.

Possibility of subscribing to certain users

[steep magnet]

Suggestion: Can edit lyrics in the final songs.

[buoyant dawn]

U can put it in the style

[untold moat]

(I mostly use custom mode) generate title and style (not random, but related to a prompt of how u might want it to be maybe), also editing lyrics while keeping the same style would be nice, this probably isn't a problem with a subscription, but when I get a good style or song or how it sings it, I can't fix it or find it again with 40 or less credits, and it sometimes skips lines, I'm not sure how to prevent that but it's annoying because it can waste credits, it also sometimes goes over the 2 minutes limit, which is easy to fix by changing the lyrics but is annoying if I'm out of credits, because it's easy to waste all of them in 1 song for me, I mean I did yesterday and today, is it possible to force it to end in different ways, like smoothly ending or abruptly? Or a regenerate button for 5 credits, would be nice to be able to set a tone (if not possible, if possible pls tell me) for a line or verse (idk if any of this has been suggested, srry if it has)

[leaden marlin]

1. Possibility to change one word in generated music
2. Possibility to use the same music style and vocalist but change lyrics
3. Possibility to choose how long the song should be (in minutes)

[quaint valley]

need ability to send in a song and it generate variants according to the style

rather than having songs spawn out pretty much all random but good although there is only a few styles I'm looking for but cannot understand how to word it

if people can compile a list of styles of the computer to analyze, it would be easier to get what people want

[thick flare]

Stem separation available in dry and wet downloadable format, would boost your active monthly paid users exponentially.

[glacial quest]

It would be awesome if we could take a song that we like and make it a "band" and generate new songs with different lyrics and different music, but the same set of "instruments" and "singers".

[still stirrup]

why is there a bird on a lot of album covers? you should reprompt your image generator

an instrumental song with the title "Hacking the System" and genre "glitchy noise core" gives me an image of a a fucking bird xd

[tepid trench]

The AI model Suno uses is called "chirp". I'm pretty sure they include this word into the image generator prompt. If you have lyrics, it will consider those and you can get things other than birds. Instrumentals though will usually end up as birds.

It'd be nice if it looked at the style or allowed regeneration of the image based on custom image prompts.

[worldly sphinx]

came here to suggest this as well

[carmine quest]

I do not get any birds for my instrumentals, but rather pictures more or less related to the music style.

[still stirrup]

If I use custom mode, then I dont get any birds btw

[glacial quest]

Another awesome feature would be to have Suno generate so that the song is a clean loop.

[ruby stag]

Would be nice to be able to choose which image to use for "whole song" from the parts it's generated from. Some images come out well while others... Are a bird.

[glacial quest]

Yeah, chrip really goes overboard on the reverb and shrill vocals

Too much overmixed vocalizer pop in the training data (and the world)...?

[humble summit]

Anyone knows if we will be able to edit the lyrics on the finished song because sometimes I end up cutting the thing and the lyrics I originally put in are not there and it makres it confusing to follow

As in the text

Not actual song lyrics

[oblique birch]

Here's my list of feature suggestions for future updates:

• Option to set BPM
• Option to set key
• Option to select chords to use in the set key
• Option to set effects (reverb, chop, distortion, arpeggio etc)
• Option to select voice gender and type (e.g. male, raw, baritone, harmonic, brightness, loudness etc)
• Option to set singing style (rap, belting, breathy, etc)
• Option to set song structure with parameters for each section, e.g. [Intro, instrumental, piano, melodic, 8 beats] or [verse 1, acapella, harmony, reverb, vocal chop]
• Option to select range (lower and upper notes)
• control lyric timing based on beat, e.g. Please[1/8] stay with[1/8] me[3/4] - the values between brackets are the beat lengths each word should be sung.
• Option to set an approximate length of the section to generate with a cost per x seconds. e.g. 30 seconds of generated audio for 5 credits.
• Create folders in the library
• Group song segments
• Download stems

[timid night]

Please allow visitors to view a playlist link without logging in or needing an account, else I cannot share my suno playlists with most of the internet

[late ginkgo]

Make that we can change picture of our sound

[sly jetty]

how to do the channel

[hybrid crane]

If you mean using discord for creating music? That doesn't work anymore. They fully switched over to the webapp via https://suno.com

[sly jetty]

OK , I got it

what does this channel do ?

[hybrid crane]

This discord if for information about Suno in general. This channel where we are now is purely for giving suggestions for Suno.

[sly jetty]

OK , thx

I found a lot of people chatting in general

[raw mist]

Ability to get the key, BPM and chords to a created work.
Thank you!

[trail slate]

Please Increase mp3 bitrate to 320. That helps so much

[chrome nova]

For in the suno-showcase channel:

Is it possible to have more details show up when people paste links here in discord?
(mostly so I don't have to click everything just to find out it's not in my style again)

Currently it just shows the url page title, e.g. "SONGTITLE | Suno" and an image.
Is it possible to show more data like style of music in here?

I do see useful metadata in the page source of the given url,
like: <meta property="og:description" content="STYLE OF MUSIC. Listen and make your own with Suno.">
maybe discord can parse that?

[tawdry wave]

Algum gpt para rearmonização musical?

Quero apresentar uma progressão de acorde, por exemplo:

[verso]
C | Am | F | G | C | Am | F | G |

E o gpt me responde criando uma (ou mais) estrutura mais complexa e interessante com diversas alternativas de caminhos harmonicos melhorando e enriquecendo a progressão de acordes.

Ou agindo como um determinado musico e fazendo a rearmonização da progressão de acordes proposta ao estilo dele.

[distant latch]

An option to recreate a song with only slight variance.

[hybrid crane]

It actually depends on the person who post there. If they post an url or a 3party platform you won't see it. But if you see a suno url, you can just click that. It will pop up another page and it will show exaclty the style of music and lyrics.

[neon harness]

At Cyanite.ai you have the ability to upload an .mp3 song for analysis. You can have the website checked in your department. If you do this for SUNO with the intention of SUNO's AI creating songs using as an example the audio that was sent to be analyzed, it simply ended for the singers. SUNO's AI will know exactly how we want it to create the songs we ask it to do. (and also an option that brings types of voices to choose which one we want in our songs, would also be welcome)
https://cyanite.ai

[visual brook]

long term, while playing with the 1 month sub is nice, i would like a publicly advertised "Enterprise grade private instance" for Suno v3, (v4 etc as they release). something that just had a yearly fee, a private end point, and does not work on credits with unlimited generation.

could even have different tiers like CPU instance vs h100 vs TPU whatever

[oblique birch]

https://chosic.com does this for free without an account or limitations... More specifically the Music Genre Finder option.

[craggy echo]

How about the ability to earn extra credits for songs made public that become popular on the platform? Or other ways to earn credits, for sign-up referrals, for example.

[cinder roost]

Hi, i don't know if enyone as already said this but, it would be cool if when we have finish multuple songs et make the full song, we can edit the lyrics and the themes, because, at the when the final song is done, every things i have whrites si in, et it make non sense, so we can make it clean and read the lyrics better. Thank for reading, if you do. 👍🏽 🙃

[chrome nova]

I like suno.com/explore , it's great for getting to know new genres.
But I would love a little more fine tuning for if you want to explore in a certain direction.
(for example, I might be in the mood for "swing like things" , but not so much for "heavy metal like things" at the moment)
For example a simple text filter for style, e.g. "swing" filtering all styles containing "swing".
Or perhaps a way to search for related styles by using some kind of relation graph using a starting genre and max distance, or a tag cloud or something similar.

And related: I would also love an explore or browse mode with more fine grained control to find songs directly.
e.g. suno.com/browse
Query options could be things like:

• genre filters (one genre, a genre substring, a genre and relation distance, or a "include related genres" checkbox)
• filter by songs the creators themselves liked (should improve general results)
• search for "topic" string matching in song title, song text, song description or creation prompt
• perhaps search by tag, if we can add tags to songs
• checkbox for sprinkling in some wildcards (just) outside filter parameters
• ordering options (e.g. newest first, or most upvotes (in recent time))
  The results could be shown as a simple table list of songs, or perhaps as a sort of dynamic playlist or queue or radio station.

For example, if you want to find songs with a genre related to swing, something about robots and dreams, you can search for genre "swing" , topic "robots dreams", check the sprinkles, and then enjoy the communities finest robot dream swing songs, with some wildcards mixed in.

[unique turtle]

Have an option that a Playlist loops once the last song is done, kinda annoying to do it manually.

[upper isle]

It would be great if subscribers could continue to receive 50 credits per day on their account.

[half jewel]

am i crazy, or is there no search functionality?

[hybrid crane]

Yes Suno is still a young company. Much is still being developed.

[chrome nova]

I would love an oversight of all the songs I liked of other people.
I Spent some hours exploring other peoples work here in discord, true gems there, but I if I don't bookmark them I don't know how to find them again 😢

Also I wouldn't mind upvoting not just a song but also creators.
Also, upvotes going to extra credits could be interesting. Or a donate (my) credits button as a "I'd like to see more of this please" button.

[visual brook]

https://suno.com/playlist/liked this should take you to a playlist of all your likes - it's babyfaced, just a few days old feature

[chrome nova]

Ooh thanks! That's brilliant, exactly what I was looking for.
Where could I have found that link? I don't see it anywhere

[visual brook]

on PC chrome, Home page, scroll down 🙂

[pseudo swallow]

A button to repeat albums. I put my working album on, and it cuts on the last song. If we get a "repeat album" button similar to the repeat song button we have we wont have to experience that pain ever again!

[glossy ginkgo]

FEATURE REQUEST: on playlist screen, add a shuffle play button to play songs in random order

[chrome nova]

Whoop, there it is 🙂 thanks!

Funny thing... I never looked at the homepage after my first visit to suno.
There's more good stuff there 🙂

[solid hound]

Add a discography section on profiles like Spotify and the ability to create albums that show up on profiles

[chrome nova]

I would really like it if we could add a private (your own notes) and public note or description with a song.
A private note for how you made it or other notes, and a public note or description where you can tell how you made it or what your idea behind it was. Or if you made your own text or generated it.

Also, comments for other people's songs would be nice. For praise, suggestions, or questions on how it was made.

[sour hawk]

I don't know if this has been said before, but an option to add section "before" in extend and not only "after". That would be game-changing for situations like when it skips the intro, etc.

[frozen lagoon]

Would be great to have significantly better audio quality. Also to be able to download the "Whole Song" as a wav file would be good, as the MP3 is very compressed with loads of artifacts etc

[oblique seal]

Would be neat if we could use ending instrumentals from a full song as a template for a new song. Would be cool to make songs that blend from each other

[neon harness]

If Cyanite wasn't just 5 songs per month that you could analyze, it would be far superior to this site. (but I also use it)

[opaque gull]

More influence on the outcome

[proven apex]

Hello! From Brazil Rio de Janeiro, I would like to suggest the option of separate tracks for download or the instrumental base separate from the voice, so that we can put our voice in place and release it on platforms

[fossil kindle]

Can we please be able to multi select songs for trash/deletion?

[hybrid crane]

You can.. just like with windows.. 😉

Click one.. hold shift and click more..

[glossy junco]

why isn't there a clearly labelled "sign in" button on the site?

seems like a pretty big omission

[lethal trail]

Would it be possible to set up the Woodhouse bot to react to messages which includes the word "chirp"

[inner jetty]

Would be nice if you could create a duet and specificy which voice was which part.

[spiral dome]

Hey there! I have a trouble, Suno refuses to generate my song despite it doesn't have any banned words etc. Can anyone tell me what does she react on???

Verse 1:
In dim light, by the table, I calmly stand,
Arms in check, my tools at hand.
Can't linger long, the road is my demand,
I'm the hunter in the dark, justice's final stand.

Pre-Chorus:
They say we're equal under the law's grand scheme,
Yet it's a fable for the fools, not this reality's regime.

Chorus:
So I set out for the hunt, to stalk the vile elite,
In their secluded halls, where foul plays repeat.
Untouched, they revel in filth, no one dares to compete,
But I'll face them head-on, ensure their defeat.

Verse 2:
The degraded world, their plaything, rich in vice,
I'll be the end of pretend, exacting the price.
To fight fire with fire, I've sharpened my own blade,
Heart steady, aim precise—my resolve never swayed.

[meager condor]

just yesterday when on the create screen I noted how the song didn't disappear when giving it a thumbs down and was grateful for the change. Today I was disappointed to see it's reverted back to being removed. It would be helpful to have the song more easily visually identifiable in the list to avoid accidental deletion of the wrong song.

[formal bear]

it is still in your library, different from delete

you can find it

[meager condor]

@tony i'm aware of that

[formal bear]

maybe we can do a phase out rather than immediately delete

[meager condor]

why change the feature then change it back so quickly

[formal bear]

cause it was a bug to change it at first place

[meager condor]

i've been meaning to ask, is there features only available on discord that isn't on the web app?

[formal bear]

lol no

web app is the go to place

[meager condor]

thanks

is there a way to get consistent original sound coming from a v2 song to v3

a search function would be great or at least a way to enter page#, it's not fun needing to mouse click 80 times.

[leaden flame]

Suno suggestion: Desperately need an 'Empty Trash' button on the Trash screen. If you want to clear out your Trash, you have to delete each individual audio track separately... Yikes! 🙂

[worldly galleon]

It would be helpful if auto-generated lyrics could optionally be extended when going to extend a track. If I'm not mistaken, right now when extending a track, you're required to bring your own lyrics. Not sure what prompt Suno is using, but it gets much better results than I do just going straight to ChatGPT.

[formal bear]

you can try to continue off v2 song

[worldly galleon]

I don't doubt that, but their prompt is obviously tailored towards writing great lyrics. Mine isn't 🙂

??? I thought this was the suggestions channel

Why even have the feature for Suno to generate lyrics then lmao

That's me

[leaden flame]

Not sure at all, but I tried it in my Suno account and also got a warning message stating it couldn't create it and that it was flagged for moderation. 😬
Maybe this part "my tools at hand." was interpreted in the wrong way. 😁

[meager condor]

it's also still remains in the create screen song list, just hidden and possible causing more confusion as to what song is being played on the list.

[weary turtle]

Can you add throat singing as a genre?

[weary turtle]

I totally agree a duet would be awesome. Especially if you could use the metatags to indicate which verse, chorus etc that you want it to change or be male/female.

[meager condor]

you can, just have to play around with styles, try using masculine and feminine first names and pronouns in different places etc caveat doesn't always work

[urban sleet]

Search in the Library would be nice. Even if only for titles.

[uneven cairn]

to include Cantonese language/vocals?

[neat trellis]

I think it is necessary to control the overall structure of the song and the mood of the paragraphs by introducing tags in lyrics, such as [Verse],[Bridge],[Chorus],and so on

[ebon willow]

Hi everyone, new here. Is anyone ever asked if it’s possible to have the song in a better audio format as Wav or AIFF ? That would be awesome for pro usage

[neat trellis]

I found in the song with lyrics of Chinese words, the pronunciation is a mix of Mandarin and Cantonese

[buoyant dawn]

Optional custom image for songs

[uneven cairn]

really!? the songs with Chinese (either simplified or traditional) I've been making was all Mandarin pronunciations none was Cantonese.

[finite plinth]

You should be able to do this with your songs now

[cerulean creek]

Could we please get chorus effect turned off as default and ask for it in our prompts? Why is extreme chorus default?

[patent axle]

is there a way to download all of my own "liked" songs at once, or can i only download them one by one?

[short solar]

suggest how about giving us the ability to do duets you know its been a thing forever

[ruby stag]

I asked this earlier and there is a FAQ about it. The TLDR is it is very difficult and requires lots of regenerations to get what you want.

[ruby stag]

One at a time afaik.

[short solar]

i see

and no way to extend with a male voice i tried it

ill stick with wanting to be able to extend the time how i want it cuts stuff off to much

745 credits to make 7 full songs so i guess i need the 10k if i want to do a lot more

[ruby stag]

I find if I am less picky it uses fewer credits... But yes. Sometimes I want just that one thing and 1000 credits later I have the song.

[short solar]

its got to sound right thats all there is to it

[stable gust]

shuffle button when I'm listening to music

[ruby stag]

Search function to find other people's songs would be nice. 🙂

[proven apex]

I'll talk again about: Having an alternative to producing 2 identical songs, 1 completes the other just the instrumental.

By adding a human voice, it becomes easier and cooler to get on streaming platforms.

It would be a surreal advance, anyone who agrees reacts for the moderators to see and respond. Thanks!👆

[proven apex]

I did it with more than 3 minutes expanding,

[primal basin]

⭐️⭐️⭐️⭐️⭐️⭐️⭐️⭐️⭐️
"Empty All Trash" option
⭐️⭐️⭐️⭐️⭐️⭐️⭐️⭐️⭐️

[frosty breach]

A warning message on click on generate while everything is blank would be nice. It's a common mistake for me after having to hide the custom mode to view past generations.

As a parody of tragic comedy, I just did it again

With smaller windows, custom mode hides it, so you'll have to disable it to see the generated songs. But then after I'm not satisfied it's common atleast for me, to tap Create again.. without looking if it's in custom or not.

[distant latch]

Could it be provided as a stem or midi format file in the future?

[tidal wharf]

Can we have a “full/not full song” filter in the library pls?

[tidal wharf]

Yea, that's super annoying. I ended up using two Suno tabs on the phone, one with the prompt and the other one where i check the generations.

[frosty breach]

Oof

[tidal wharf]

It's more convenient than it sounds lol

[fathom crystal]

Extending songs is nice, but it would also be quite useful to have some basic editing and trimming tools to stitch different versions together - sometimes a verse will come out great, but it flubs the chorus, and vice versa.

On a similar note, sometimes songs don't end neatly. It would be good to have some basic transitions to fade out the song, or end it abruptly on a record scratch, etc.

[grand haven]

Hey

@fresh shard_

why isn't my music showing in the 'New' section? I generated it a few days ago but just made it Public today. Thanks!

[potent ferry]

Wild guess, maybe because there are thousands of song made every minute and the "new" site can only show so many. I was lucky to see one of my songs on "new", once.

...but then again, that's just a guess.

[dusty raft]

Is it possible to automatically add all new songs to a specific (possibly new) playlist? Then everything is better organized and sorted.
If not that would be a suggestion from my side

[unborn token]

The dream feature. I don't know it it's even possible and proly not gonna happen soon, but : a way to **enhance **the quality of already generated music. like from V1/V2 and even what we make now on V3, some have white noise, dirty sounds, and overall sometimes a lower quality, specially some genres like rock/metal. so I way to improve those.

[dusty raft]

V3-Alpha is reffered to as a better version than the now V3 :/

[patent axle]

just found out that (on windows) you can click the "..." then type "D" "enter" "enter" so it downloads the audio. at least saves a couple clicks and mouse movements 😉 if there is a shortcut for the "..." i haven found it yet

[leaden marlin]

1st suggestion: OPTION to CHANGE text and music remain the same. Sometimes we do typo or we eant to omprove text, sometimes Suno make error - but music is perfect. Next rendering might get lyrics better but music is bad ... so we end up making 100's of attempts. That uses up server any out time. You might want maximise profit, I get it. But en the end we are saturated of frustration and we'll eave for other service. There is no point loosing money and time for something that does not return even slightest fun out of it.
2nd suggestion: try to give exact manual what works and what not. If I write [instrumental] in the begging of lyrics I do expect song does not start with verse. Make a list of commands that are actually understood with AI. If I say [Slovenian language, CLEAR solo VOCAL], that I do expect it is one person singing only not choir of harsh guttural raucous voice sunken under acid music bed. It is expected that we get manual what options are available and what is pointless to write. It seems SUNO often does the opposite of those instructions. And language sometimes is read perfectly, next attempt it would have some foreign accent ... etc.
3th suggestion: make possible to extend music BACK from the time. When perfect suggestion is rendered, but it cut off 1st line of verse fo some reason. It would be perfect to recreate song back to make decent intro. Also SUNO do many times does start song with lyrics frm 1st moment, without music intro.
4th suggestion: make possible to mark what to use in whole song. Sometimes in one take, only part is superb. let say CHORUS, ... but other parts are bad. So make posible to mark what is good. And to be included in final song. (I download parts and assemble them in software for now, but that can not be published on SUNO AI site).

[hybrid crane]

Most of you suggestions (1, 3 and 4) can be done in the future with the upcoming "paint-in" feature. And for your 2nd suggestion, it's not that they Devs are witholding us any information. They know what music they have used for teaching the model (genres and lyrics) but that doesn't mean they fully know how to talk with it. They are learning with us. There is no manual, this is not traditional software.

[leaden marlin]

I am happy 1,3 and 4 will be posbible.

I am a developer of web apps. There is always "manual" possible or "instructions": "Try this, try that ... do not bother with that ...".
DEVS should know if "instrumental break" will do anything or not. If this is not the case as you say, than it appears as if they do not know what the are doing; they did something and learn all together what works or not. And they are charging than for this development and debugging from our side? I learned that [verse], [chorus], etc in V3 does little to none. Lyrics is used how V3 wants, regardless if just written in one long line without any breaks or lines. If I happen to nail structure of the song the AI just has in it's "mind", it will come perfect, otherwise will be trash (again) regardless if I put ton of instructions and [comments] within. The problem is: every attempt is something completelly else and we can never know how AI will proces lyrics this time. So adapting lyrics is dubious tactic again. I struggle to try to re-write lyrics in order to make it spell correctly, but it is endless tail chasing.

I see Suno good for IDEA, than you go make that song traditional way. Also to make quick 2 minute fun part for friends. V3 really push to make complete song in 2 minutes. But here already we mostly need a lot of attempts to get usable vocal. English is mostly good, but my language is good only in 10% of attempts.

[royal venture]

I joined this Discord channel hoping there would be a list of [commands] that are actually useful in song generation, but is there no consensus on what actually works or what doesn't?

[leaden marlin]

EXACTLY ... my point above. We shoot in the dark with attempts (and paid credits) and hope to get something.

[royal venture]

hm I see thanks. I just saw this song trending on the site and it has plenty of [commands] which seem very logical but then it sounds like it's a roll of dice 🙂 : https://suno.com/song/93288786-8d6a-4e6a-95fd-4e3ba1aca02c

[young valve]

absolutely a must -> continuation from a different segment

than originally created

say I experiment and have 5 different versions

I do want to be able to choose the segments that will be in the final song

next -> editor to precisely edit and cut, paste, etc

the first one should be easy

10 minutes for a programmer

I would also like to continue with a voice that I find interesting ... so maybe let me choose that parameter for the next song

that's #3

#4 I want to choose BPM

[young valve]

#5 finally something that can make a difference -> allow upscaling the song to higher resolution -> you could charge $10 for this and you'd make a bunch of money easily

a couple of choices of coures ... if someone wants a master tape quality then $10, slightly less quality of course less

in fact without this you'll be only a hobby that will eventually fade out

[old relic]

repeating the answer you already got: this is not traditional software. There is no way to know how the neural network will react or what is that will work or not ahead of time. Devs know which annotations went into the training data, but not what the network learned. So, it is not possible to write an exact guide of what works or not. Devs are probably learning from us in this respect

[young valve]

true, the model is an artist that is freely creating around what you want

[heavy jacinth]

if it's possible, i wish there was a feature that would remove the vocal of a generated song and give you only the instrumental

[leaden marlin]

correction: the model is an artist that is freely creating around what HE wants 🙂

• it skipps likens or whole verses
• it ignores [verse] and [chorus]
• it ignores [istrumental], [music intro] etcc and sqeeze all lyrics in 1st 45 seconds than leve minute or more instrumental, but start is with lyrics without music bed ...
• etc

[leaden marlin]

and that is why I am against all that "machine learning" (e.g. "AI") because we have absolutely no control over it! People developed our brain over 1 million years from apes, but here software sample couple of thousands of songs and we leave complete open hands to it how it will use that data?!? We must limit and guide the models! If "[male singer] is the instruction than model must not use female voice, period! Are WE the masters or there is machines? And to achieve this the "software" for this so called AI should be more complex. Not just sampling data and than hoping it'll recreate something, we should articullate structure in advance by giving model x-channels and clear instructions. So we should have models for separate tracks and vocal! So vocal is always vocal only and than is no problem changing vocal and music remains. And follow instructions how vocal should sound. Same for instruments. Sample intruments and music knowledge not (just) finished works and hope it'll come out good. I am now burned like 350 credits on EDM 1990s song that constantly pushes some acid synth in the mix, hopeless, and make voice like aliens gaggling some petrol. 😦

[wraith adder]

Could you please add a max time option to be more then 2 min? It is really annoying to have to extend and piece together parts especially when you use your own lyrics. You can have the user choose the max time and for every 2 extra minutes you can charge 10 credits.

[leaden marlin]

the most annoying is when SUNO skipps lyrics ... burned hundreds of credits on repeating attempts where line or whole verse isskipped and it stars with lyrics further down ... clearly not what WE want.

[wraith adder]

After the style of music put male. That works for me most of the time.

[old relic]

I don't think this is the right channel for this discussion. But the answer is simpler haha: this is how current technology works. This is the best we know how to do right now. If you don't fund it suiting your needs, you can just not use it, or work towards improving it 🙂

[leaden marlin]

sometimes ... in my experience 50% time this works. LEt say I make 20 versions of same prompt 10 will be opposite of what Iistructed. Also fo lyrics ... for last hour I struggle with perfect extension. 10 attempts a line of verse is skipped. Than suddenly it works, but language is completelly off. Than something else ... I'm done when I burn my credits here. Waste of time. the 50 free per day is enough to make some fun. Buying 2500 and burning them in a day is not fun. 😦

[molten robin]

It would be useful if we could keep the original track image, for the extended song... because the extended songs image tends to be the image for the full song? (and the extended songs image (if instrumental) tends to be a bunch of instruments!

[quartz steppe]

one thing that I figured out in hard way is that prompt doesn't really matter, what really matters are tags that you see attached to generated tracks
only when I began simply listing all tags, instead of describing anything, I started to get results, close to what I am aiming for, more frequently
can't say I got what I wanted after burning around 2000 credits, but I am still trying
and what really bothers me with that process is that I can't provide proper feedback, just these likes/dislikes
model training goes painfully slow because of this

[mortal shell]

So, check it, I'm keepin' it tight, tryna stick to just 100 creds a day, ya feel me? Hopin' to cook up a dope full track with that. But if I hit that 50 cred point and ain't none of the jams hittin' right, I'm out for the day. Gotta switch it up to the freebie account, testin' out prompts and song structure, see what's really poppin' off.

[visual brook]

i visit New and Trending when my gens don't pop off

[leaden marlin]

can you please suggest what you had in mind "tags that you see attached to generated tracks". Where I can see those TAGS, so i can use them?
(but yeah, similar to generative fill in Photoshop. The more you describe the less you get out. Describing "black large steam engine" gives "blue red Thomas the tank toy image ... ":(

[quartz steppe]

well, I just called it that way, thought it would be obvious
you can see them under song title, after it gets generated

so, if I, for instance, specify a "drum and bass" track, and I don't see it in tags
then, basically, generated track will have a random genre

[leaden marlin]

yeah, I get that. But those "tags" are prompts that we put in in "style of music" window. I hoped we can see some tags somewhere else, so we know what to use.

[wooden ridge]

~ Please make it so that songs move from your library into the playlist, instead of just filtering it. I want my library clean without needing to 'delete' them.

[visual brook]

also, have to somehow watch the video if the tags change between extends

[quartz steppe]

if prompt is too long, or has some phrases or whatever
then these tags are added randomly in every generation
and will ignore most of the prompt, because it contains like 4-5 tags in such case

then I begun simply adding these tags to the prompt, and I get all of them attached to tracks
it is very confusing
it would make more sense to have ability to list all tags directly
rather than writing a prompt, that is being reduced to tags by the system anyway, and I don't control the process

[quartz steppe]

"then I begun simply adding these tags to the prompt, and I get all of them attached to tracks"
just as I mentioned it, Suno threw away almost everything, leaving only 3 tags
and as a result, wrong genre, wrong everything, another bunch of credits burned for nothing )

[hybrid crane]

This is turning into a chat channel, while it is meant as suggestions for the Suno team. Please move your chat either to #💬┃general-chat or #📚┃suno-school

[severe glen]

Ability to edit same song to make changes to lyrics... 😁

[scenic oracle]

Would be great if WAV format is also available for downloading tracks

[bitter tapir]

I guess this might accomplish the same kind of effect but I also think some kind of seed function could open some great possibilities.

[calm abyss]

when are we gonna see a Suno App?

[cerulean creek]

[final chorus] and my vocalist sang " final chorus" LOL

[modest cobalt]

ability to add your own vocals for deeplearning, ability to add your own instrumental under the custom track generation. adding more payment methods, via paypal, etc. The current payment methods are quite limited, i only have a bank account and im stuck to generate 10 songs a day which is a bummer because i would defo buy a subscribtion.

[viscid cloak]

I will make a big Suggestion:

A Copyright Discord Channel, where we can ping Song Stealers, Ping people who got stolen songs etc and help People resolving the issue.
I already claimed back my stuff and it worked and maybe i can help others with that.
It may be also a good way to warn people, if they make their stuff public etc.
Maybe discussions about the whole topic itself could happen there

[mortal shell]

For songs created within the free plan account, maybe add a "Made with Suno" similar to what Winamp did back in the day. If free users want it removed, they can susbcribe and devs can take that out
https://youtu.be/oQid2jSU7Ww?si=I0sf-gPNPD4cSw05

[viscid cloak]

Songs created while not subbed, stay that way.
Otherwise the suggestion is good to see which you can use and which not, it would help some people to keep track of what they can use commercial and what not.

[craggy echo]

You might find this helpful: https://www.suno.wiki

[cerulean peak]

Hey, is it possible to pay via paypal in the future. In germany kreditcard is uncommon

[half sail]

PAY YOUR MODS - thank you 😄

[formal bear]

agreed, controability and reproducability would be great

[grim pagoda]

A "seed" set as any image AI would be appreciate, so if we like a seed we can use for other lyrics

[glossy ginkgo]

Suggestion: Make song pages have a tip feature where you can tip 5 credits from your account to another Suno creator so they can make more songs.

It'd be a good way to spend those last 5 credits

[soft frigate]

A way to get the stem of the song you really like. And connect them

[rain radish]

WEEKLY SONG COMPETITIONS!!!!

OR EVEN DAILY!!! =DDD

[wooden ridge]

I don't really want to compare apples to oranges here, but I was playing around with udio (which ive had no luck with really) but I saw on their remix button it pulled up this.
Something similar would be fantastic.

Just some sort of way to keep same vocalist, or vibe, or something would be fantastic. We need more control over things ~

[quiet swan]

get a proper feedback management platform (like canny.io)

[jagged fable]

In future versions I'd like to see "style of music" prompting meta eventually evolve into being able to handle more direct/specific commands rather than just comma'd lists.

For example, have it be able to understand something like: smooth jazz verse with complex chords that transitions into chaotic prog metal chorus with a xylophone solo, the key changes 3 times throughout

Rather than: smooth jazz, complex chords, transforming, chaotic, progressive metal, key changes, xylophone and then just hoping that the list is interpreted the way you want it.

[viscid cloak]

I saw that you guys do hidden audio detection for Suno songs, could it be possible to get that for the Song Owners too? So you could use that, to justify its your track.

[spiral dome]

Strange, later I made it work, but not sure how exactly that happened...

[lavish meadow]

what is a more descriptive style than edm for this https://suno.com/song/990ae77a-0f39-4c4b-971c-87c304b8c4fe

[leaden flame]

Maybe a Suno site moderator later approved those lyrics.

[frosty breach]

Custom header for profiles might be neat. They're a bit barren to look at.

[vague folio]

it would be cool if i could enter my desired BPM, as well as how many bars of intro/outro i want. i dj/mix and would love to throw some of these songs in.

sometimes banger songs just cut off and i have to grab an earlier part to loop

[wooden ridge]

~ Show linked songs (different parts) under the completed song so we can easily find them.

[rain radish]

I wanna hear that song

[timid night]

Just tried it again and it looks like you guys were able to implement this in only a day! thank you! now I can link people to decibelcortex.fm and it will forward them to my suno playlist woohoo, thanks devs

[formal bear]

It's not like we don't have buttons to push around, we just don't want to give them to you folks unless we are sure they kinda works

[wooden ridge]

No thats understandable. I'm hoping i didn't come across wrong here, I absolutely love suno. Definitely looking forward to anything you guys have cooking.

[grim cedar]

I think it's been suggested before but if each 'project' had its own section with all the parts and only the final product was put into your library (maybe a toggle for finished, similar to public) then the library would be so much neater and all the parts would be grouped together, it would be infinitely better from a user perspective.

Right now my library is just hundreds of pages of parts and i'm never going to be able to use it effectively.

[young valve]

you should have a feature list where people vote

as I imagine the dev resources are limited

and it takes time to create new features

and some are impossible today

...

https://suno.com/explore - it is too complex - I want a tree where I can see all possibilities quickly of what genres interest me

not 100k of options where I can't find anything

[hybrid crane]

See here how it works: https://youtu.be/kB3W5xLza_k?si=Lnohi_jjWApj3aER

[young valve]

I've been using it for 20 minutes so I think I comprehend how it works

it's a chaos ...

good for showing off

but not practical

[hybrid crane]

I think it works great.. You can reuse the prompt, see the lyrics..

My videos is just 1.24 so it cannot hurt to watch

This is the wrong channel for that. This channel is meant for suggestion to the Company. Not for feedback on your music. Please post that in #🐣┃suno-showcase

[lunar dragon]

cool thanks

[kindred minnow]

Hello everyone,

I'm trying to create music, with 2 voices (a woman and a man) to make a Duo.
For example, the male voice sings one verse, and the female voice another, and in the chorus he sings at the same time!
I tried everything [female voice] for example with every possible combination.

Hence my question, is it possible to make this kind of music?

[hybrid crane]

This channel is for suggestions to the developers. Not technical help with making songs. That's the #📚┃suno-school for, and also #💬┃general-chat

[kindred minnow]

Sorry,
But it can be as a suggestion, if not implemented.
It's to add a kind of example [Voice:duo] [Voice:Female] in front of a chorus or a verse.

but I will ask my question in this channel thank you.

[hybrid crane]

But we have an answer on this.. There are duets made and we know how

So please move this to the right channel. Thank you

[kindred minnow]

ho thanks !!!

[light bramble]

It would be awesom to have the ability to specify notes to syllables: for example the phrase: I want this specitfied - [C]I [D]want [E]this [F]spe- [G]ci- [H]fi- [A]ed? and with the "-" symbol you could help the program understand that it is a continous word and not just sounds separately

[glossy ginkgo]

Suggestion: Update channel description for #😍┃v1-songs

[fringe crater]

Wallahi i beg please make the suno ai recognize tags or acknowledge style of music changes in second part during the generation where for example you either add a tag directly to the text where you can specify what kind of genre it should use or specify a part either directly in the text... or while generating an actual second part of the song allow us to change the genres and switch up the beat , pace etc... especially if you make it recognize [Fake Bass Drop] [Bass Drop] properly it will allow for some insane genre mixes , switch ups , fake drops etc...

**`>>Techno Genre Start;
text prompts , etc...

Techno Genre End;

Hardstyle Genre Start;
text prompts , etc...
Hardstyle Genre End;`**

Either like this or like this

**`-Lyrics Custom Mode
[Part One Start]
lyrics etc..
[Part One End]

[Part Two Start]
lyrics etc..
[Part Two End]

-Style of Music
[Part One] techno genre 150bpm
[Part Two] hardstyle genre 180bpm`**

I swear to god if this gets implemented i will rave in my bedroom for 5 days till my heart gives out

[pine basalt]

I'd like to suggest the ability to swap the art generated for a song to one of the other ones you've generated, or at least in the chain you've generated. See I have a song, that I have perfect, lyrics and melody are all where they need to be. But the art is off, it wasn't before when I generated part 1 or 2, but by part 4 it swaped the art to something not related to the song. So I'd like the ability to swap to one of these earlier arts that were generated.

I get not being able to add your own for legal reasons, but at least let me swap back to one of the arts that was more fitting to my song.

Edit: I dont want to try and regenerate the whole section or track because it will change the sound, i got that perfect I dont want to lose that

[hallow frost]

CAN you guys make an EASY ACCESSIBLE list of what will GET FLAGGED FOR MODERATION? I am just trying to make some rap music I have no idea whats triggering the moderation

[hybrid crane]

It's not words in general. It's about the context of the story. You can paste your lyrics in chatgpt and ask why it is flagged for moderation.

[hallow frost]

so is like a diss track a context that might trigger that?

talking bad about someone etc? or what is it?

[hybrid crane]

You will find out if you follow my advice 😉

[hallow frost]

ok

nice i tried to tell chat gpt my lyrics and it got flagged

guess I should just keep em in my back pocket, i might be the last human artist on the internet in 10 years

[hybrid crane]

but it still will tell you.. Usually

[hallow frost]

no it didnt

just "violates terms of service"

[hybrid crane]

That is the first time hear that.. So that it is absolutely too graphic

It's not just a few words then

[hallow frost]

well i dunno it's not outside the boundaries of what major labels have published in the past

but yes it is memphis phonk

[hybrid crane]

That's not the point. Suno is not a guitar but a service and has rules.

And sadly they are PG13 atm

[hallow frost]

ok thats why I wanted to know what the lines were

so i can color inside them

(riiiiight up to the edge)

[hybrid crane]

I get that.. and my advise always worked..

😬

They are still developing thought.. hopefully they will change it in the near future.

[austere sail]

How can you tell if you have flags against your acct?

also trying to get strictly the instrumental for a generated track, is that possible (on a subscription plan perhaps?) or do I have to just play with the extend option till the cows come home lol

[limpid fable]

I'd like a search feature to find songs made by others

[open saffron]

Personally, I would love to be able to edit the lyrics once I have the complete song, many times when part two or 3 remains the previous lyrics in which there was an error and I would like to leave a better finish

i mean in this part

[hallow frost]

when you remake it you can retype the lyrics

i have a 13 minute song so far

you can change the style to like to get different vocals too

[merry thunder]

make prompt charts!

[royal cipher]

Can suno currently provide API service?

Can Suno provide API services? For some companies or enterprises

[merry thunder]

yes i would love to have an API so i could make my own frontend and/or pass the music automatically to further processing or do other creative stuff with it ... that would be really nice

[royal cipher]

You are an official staff member of Suno

[merry thunder]

no

[scenic lichen]

1. We could really use a structured lyrics input or "general song structure" where you can pre-set how the lyrics are generated (intro, verse, chorus, verse, chorus, bridge, chorus or whatever you select) up front because it's really difficult to get the song structure right and then add an ending/outro fade out at the end. 2. It would be nice to be able to comment under songs or see who liked your post so you can see their work also. 3. If we could have a DAW style waveform or some kind of visual editing interface to be able to say "hey more drums here" or "big climax here" or "drop here" or "i like this more of that here" such that we can interact with the output like a "mixer" to better tune what we have or generate segments, or even other instruments in different tracks that would be a great next step. 4. Often I produce multiple rolls that end up sounding similar but have different elements i want to combine -- a "merge" feature would be great to reduce the junk rolls that clog up your library. 5. If you don't like a roll then just refund the credits for that roll and auto-trash it -- there is really no purpose for those and deleting is kind of annoying. Also if you get a bunch of junk you don't like may as well let us have the credits back to improve the input and output of the system.

[leaden marlin]

Yesterday I was suggested that SUNO is AI and it can't give a "manual" or "how to use" suggestions because it is AI and no one knows how will react to music prompt. Than I read that it is a "artist that creates around what we want". I suggested that DEVS should at least make some guidance what works and what not. Not possible, the reply. Well I went searching, and found UDIO! Not only the give FAQ and tutorial, they also DO follow the style in prompts. With SUNO user have almost no idea what the outcome will be. This is A problem! I can't write lyrics to SUNO since I can't know how many lines "it" will need for verse/chorus for particular style. I render and figure out it reads 4 lines par section. So 4 lines + 4 lines of verse, than it wants to sing chorus with 4 lines. I it happened my lyrics followed that, song will be superb. If my lyrics has 3 or 5 lines, the song will be total mess. OK, got it, I'll adapt lyrics for the style, right? Nope! Next attempt of rendering same prompts will result in completely different song and possible style. So this time, the song will expect 6 lines of verse and 6 lines of chorus after that. I burned through 2500 credits in no time, fiddling with this, also on a way 3Daizy shown in it's videos. SUNO is way to unpredictable and with almost all but few style prompts it is all around the place. Doing 20 versions the outcome will be vastly different in style, feel, mood ... unpredictable to the point of unusable for serious usage (other than 5 minutes of fun). But few styles are consistent (in style) and creating within that is pleasure because style remain same, recreation of version remain in same ballpark, and I can adapt lyrics to the point it finally meet expectation of style. And some other styles are (sampled on too little samples) are so rigid, nothing helps to create pleasant outcome. I learned that forcing a lot of versions only produce trash. First few attempts are OK, than all sorts of halucination and lyrics skipping.

[scenic lichen]

also it would be cool to have one song in our profile that we select to be put on showcase so that everybody has a chance to be rotated onto a section of the home page with their best or favorite song -- pretty tough to search for anything or find anyone yet on the app...

[leaden marlin]

So in the essence (TL;DR): it seems others can produce AI to work as expected. Give FAQ and detailed tutorial how to create music more successfully. Than prompts of style really sound like desired outcome. I had struggle to recreate proper 1995 EDM song with SUNO. It gives some generic acid outcome that sounds like nothing real. When trying this in UDIO, not only it is instantly there, it also is difference when 1990s or 1998 is used. So it could be done.

[scenic lichen]

it's not that bad https://suno.com/@oboloman i came up with some amazing things --- you have to roll your initial prompt maybe ten times or so to find an intro with a vibe or feel that you like, then you have to extend from where you want to change it up. If you like a vibe but don't like the voice or elements, extend from 1 second and it will generate something close to that with a different feel you can put in the prompt. Yes it takes many rolls and many credits to get good drops but that's their business model right now. Listen to some of my stuff and it's mind blowingly good to me anyways.. I'm very happy with it

[leaden marlin]

It seems essential goal is to burn credits. And what is 30 versions ALL skip 1st line of verse, or 1st verse completelly? I really cursed like a pirate yesterday, burning 700 credits on a song, than deleted everything. Did instrumental intro (since SUNO refuse to start song with music/beat). Than extension ... skipped lyrics. Or make 30 second extension ... I loose interest when something is harder than just LAY MY OWN MUSIC in real software and is done.

[leaden marlin]

Listened ... GREAT!

• how many credits average per your song? I have no problem to PAY, but my time and frustration is crucial!

• do models sing your language as expected? My Slavic language is sang all WRONG in 90% of time. You seem writing in Hindi. How you would feel if it would be sang with Chinese accent?

• in 60% voices in my songs get unusable trash ... I am hard ti get CLEAN vocal not to be all meshed, sunken in music, harsh ...

• WHERE you got those prompts ? I hate there is n FAQ or suggestion what to use or what it works.

[scenic lichen]

social media and ai is based on gambling model. If you got a banger every time you clicked, they wouldn't make money even if they could do that. I think the best use for this is to generate ideas that you would never be able to come up with or produce on your own -- my own music is similar, but nowhere near this kind of production and language flexibility. there's something magical about the black box spitting out something just way outside your wheelhouse and blowing your mind. It's exciting. It is frustrating at times and some lines of extensions end up dead or unusable or wierd. But the handful that came through were WILD if you get lucky and persist. I played around with similar stuff on an AI art generator that was free and spent hours and hours getting something just right with free rolls. So when you have to pay for rolls (and it took me oh, 3k rolls to get say 5 great songs/ideas thats really not bad. it comes down to luck and technique and fiddling with prompts. On the hindi part -- i dont speak hindi so the reason why i used it was 1. I love bollywood songs and i want that influence. 2. The english lyrics tend to be wierd and i have no expectations for hindi and how it should sound it's just phonemes in my ear that could mean anything. English lyrics are best when simple and trite (love me baby, all night, til the morning light etc.) ... I have no idea what's going to come out in terms of voice or singing (a girl's voice for masculine lyrics and vice versa, group singing or harmony when i want a solo, and so on) -- you have to take what you are given and then re-roll it to say "enhance or improve on this" -- you have to play around with prompts and know your influences without using artist names. That's the hard part. There's no guide bacause it's a random generator.

[subtle pier]

It would be beneficial if the singing voice could be influenced in some way. The quality of the voice is poor and lacks variety. It's sound like 100 singers were put together in a bucket.

[leaden marlin]

exactly!!

[scenic lichen]

i suggested a DAW future module that would allow exactly that, where everything had tracks and you could pinpoint reroll things, change cadence etc. That's a long ways away. What I hear is that the ai turns waveforms into "pixels" and then shuffles them in like all the music in the world went into a woodchipper and it spits out particle board. So right now you just have to see what comes out and then work with what it gives you. I've noticed also that sometimes prompt elements will show up 5 or 6 rolls later on, so its really designed to require a lot of rolling to get something close, then extending and re-rolling extensions to lock in the parts you want. Not a science by any means, but it's a start. See it as fun exploration rather than a precision tool.

[leaden marlin]

Suggestion to SUNO -> check UDIO!
It seems it answers to all my concerns:

• detailed instructions, tutorial, FAQ, style suggestions (what might work)
• 4 placements: I can make good mood than add intro and instrumental sections (before or after). I can create outro for the end
• styles are acurate and consistent
• voices are clear and comprehensible (are able to understand them, unlikely with SUNO where most what I done with 2500 credits were not clear and hard to understand singing)
• 33 second sections are fast to render and long enough for particular section.
• extensions are numbered in v1.1.2.3, so very logical. Song lenght visible.
• If I want [instrumental] section and than [chorus] at the end of that, it is respected. Not like with SUNO, which mostly ignores that and start lyrics 1st, than leave all the time instrumentalk at the end.

Is it perfect? NO! But it is welcome competition which will surely improve SUNO as well.

[outer falcon]

Right now my biggest want for Suno is at least 3 minute base song length. 2 minutes is too short. Songs really need more verses. Also the Voices need to sound a little more natural.

[spiral acorn]

Since so many people have multiple pro and premium accounts can we make it at some point so this can easily be achieved on a single account?

Some kind of slider or input on the subscription page where you choose how much credits you want, say between 10,000 and 50,000 or something like that. I know you can email for a custom plan, but I get the impression this is for even larger amounts. Also if the custom amount was modifiable by the user they could easily resubscribe at a slightly different amount when they need to make adjustments, time to time.

Btw, the playlist pinning on the profile page is awesome 🤩

[leaden marlin]

Suno should totally add an option to download/generate an instrumental version of any generated song that has vocals. I find that some of the stuff I generate has unironically banger instrumentals, but I can't exactly regenerate the exact instrumental.

[viscid cloak]

You know, that having multiple Accounts is against ToS?

[spiral acorn]

It's not against the TOS, what you're saying is in relation to the "free tier". It clearly says "you can't have multiple accounts for the purpose of taking advantage of the free tier". I'll find it for you, one sec

This is has been verified by the mods multiple times too (and some mods have multiple accounts)

[viscid cloak]

Thanks for the clarify, i was outdated

[spiral acorn]

Here it is:

You may not create more than one account to benefit from the free tier of the Service. If we believe you are not using the free tier in good faith, we may stop providing access to the Service.

[past briar]

How long has it been possible to set playlists public?

[ivory remnant]

/chrip

[zinc hazel]

After using Suno with premium credits, there are some things I have to say, the really got me.

1.) When I paid for an application, I can and should expect a sort of manual. It feels like I have to try and error, what I can exept as a test user, but when I give you money you should deliver a clear documentatiion. (e.g.: complete list of implelmentet genres and sub genre, a list of working promts and how to instruct the ai the best way.

2.) What drives me nuts: As premium member the song time is limited to 2 minutes. That would be okay, if the ai would be able to provide me the same chorus and pre-chorus if I extend, but it doesen't. If your song is longer than 2 minutes, you get not the same chorus meldody as the chorus before. So if it not possible that the ai understands, that the different chorus in a song are the same, the time of the first promt should be more than 2 minutes to get this. (I lost so many credits, trying to get the same chorus the second time in an extension, that's not exceptionable!)

3.) Same with the custom mode promt field. As a premium member this should not be limited. I know it would be senseless, if the point before is not implemented (song time limitation), but again, as as a premium member the only upside I pay for being able to use the tracks commercially and have some more credits, but having the same limitations as a free member is kind of frustrating.

[hybrid crane]

1: The Suno team does not have a manual for the AI because they cannot precisely predict how the AI will interpret different inputs. They understand the data and training given to the model but translating this into a detailed manual that dictates exact outcomes for every possible prompt is currently beyond their capability due to the inherent unpredictability of AI responses. For a "taste" of genres Suno can handle, see my video: https://youtu.be/kB3W5xLza_k?si=2jcPjdSX8nwGPB8R

2: It's important to note that while maintaining consistent choruses can be challenging, many users successfully achieve this by carefully managing their song structures and lyrics. If the lyrics are slightly off or the continuation point is not optimally chosen, the AI might not track the intended musical theme correctly. A helpful tip is to create an intermediate song part which can act as a bridge, potentially aiding the AI in keeping a longer memory and thus better chorus consistency.

3: I understand your frustration regarding the limitations in the custom mode prompt field for Premium members. The constraints in this area are not intended to limit creativity but rather reflect the current technical capabilities of the model. Although you enjoy the benefits of commercial use and additional credits as a Premium member, the AI’s processing capacity is the same across all user levels, ensuring stability and consistent performance.

I've got another another video where I show my own workflow: https://youtu.be/lj8utBd_IUM?si=Y2p_hLW9yvGeVaj1

[echo rapids]

Can we get a random tiktok like suno scroll through random suno feature?

[spiral acorn]

I think this was added only in the last 24 hours or so. It's a great feature!!!

Makes it easy to group your music by "genre" or have "albums", "EPs" etc 🍻

[fickle hollow]

Dear TEAMs, Moderators, Guides.
(All of these follow long sentences are machine translated, so I apologize if I did not convey the intent. If you do not understand the intent or content, please ignore it.)
And no reply is necessary. However, if you are interested, I would appreciate it if you could add a reaction.
First, I wasn't sure if I could even post this question and suggestion here, but it seems to be the only place, so here it is.
Using chirp has been already retired on Discord. (Because I think chirp is renamed to suno and only use on webapp.)
But, with that in mind, I have last questions and suggestions.
I have temporarily returned to the server for this purpose, but since I am in and out so often I think I deserve to be banned from here.
After this, I will refrain from post any suggestions or comments.
I will first post and reply that post after a short pause, but please forgive me if this is a series of reply.
There may already be the same post, but please forgive me if there are the same post, as it is not in thread format. and is hard to find them.

[fickle hollow]

1. Why TEAMs/moderators/Guides don't rename chirp-technical to suno-technical? When we first joined this Discord, there was (and probably still is) a guide to making music with Chirp. When we looked up where Chirp was, we found only chirp-technical, so we probably executed the chirp command there.

[fickle hollow]

2. Does "liked-clips-feed channels" work now on? I think this channel to remove or archive, If already this chat doesn't work.(I think this channel is not working now, because chirp-beta channel already does not show on discord.)

[fickle hollow]

3. Why does not write topics on each channels. I've confused and don't know which channel used what I should be talking about. (I think the topic should have a brief description of what will be discussed in that chat.)

[fickle hollow]

4. Why is the Server Guide out of date? I can't figure out how to use each of the documents to see, I can't load "How to Make Music with Chirp" in the first place, and other links may also show 404 not found.
   And no matter where you look, there is a difference between the original link and the linked description.

[fickle hollow]

5. If the TEAMs/moderators/Guides have removed the language-specific generate channels, then I think the showcase channel should be split by language. (I think this is not so important though, as each person may post in the language-specific chat.)

[fickle hollow]

6. Does "THEMED GENERATION CHANNELS" work now on? I think this channel to remove or archive like question No2, If already these channels don't work.

[fickle hollow]

7. explanation of the language-specific channels. It says "Chat with other creators who share your primary language" but I don't understand the explanation. Does this mean that I can post to any channel in any country in my primary language? (For example, if a Chinese person or not Chinese person posts in a Chinese chat in the Russian language, will a Chinese person who understands Russian continue to chat there?)

[fickle hollow]

8. In questions No.2 and No.6 you confirmed why you did not delete the channel (chat), but if you did not delete it, why did you delete the other Generate and language-specific Suggenstions channels (chat), etc.?

Thank you for reading.

[spiral acorn]

Yeah, at this stage you can follow user created guides and videos like 3Daizy and others have created. Over time there will be more and more of these! Watching how other people use Suno helps big time. You can pick up a lot just by watching how other people use it. Maybe eventually they will include videos like these as part of the official Suno guides.

This is all unfolding before our eyes and things change, so if Suno invest in writing out specifics in a manual they probably want to make sure it's "the latest approach" and also not too "limiting" (as there is a lot of flexibility here). Suno is very open-ended in the ways it can be used.

This community guide is also worth reading: https://www.suno.wiki/

I'd say the 2 minute thing is more of a technical thing than a "limit". It's the ideal amount that works for Suno's current approach.

[tacit mesa]

How can I make two different male and female voices for [verse] and [chorus] in your application?

[tepid trench]

"in-painting" has been mentioned as a likely upcoming feature, and the question of vocal removal or isolation comes up quite a lot. Demuxers always cause obvious dips in the instrumental. It'd be good to be able to e.g. remove the vocals and then regenerate the instrumental parts where the vocals were to fill in the missing detail. Unsure how feasible this is but demux by itself is imperfect so having a way to patch things up could help there.

[outer obsidian]

We need a search song button on the Library

[karmic sigil]

It might be easier to code the generation to be separate in the background and then give the option in the ... menu to get the instrumental, vocals, or both.

I think the algorithm needs to be taught what Duet and A Capella mean as well...

[lunar dragon]

it iwould be cool if there was a little text box to have a little description of the song or its intention to add onto a finished song, maybe placed after the text that shows the prompt used

[winter crag]

Can we get an Ai radio station with all the best songs? I'd listen

[merry thunder]

i wouldnt say that thats the "business model" ... its the AI and in this state of development of the AI, i think its technically not really possible to give more control over the UI. maybe they will release more and more features by the time

[zinc hazel]

Thank you for your reply. Regarding to song strucktur to keep the ai wise enought to notice that the chorus parts should be the same is a understandable claim, but to be honest, this can be done only for music over 130 bpm. If you got a ballad, or smooth song, the structur requires hard cuts out of thet song and shorten the song. So suno at this stage can be use good for 2 minutes short stuff, thats a bit disappointing. Would be less a problem, if the expensions would not be limited to 1 min, but maybe 2 min, or, as mentioned, if the lenght will be choosen free by the ai after the structure is done. The ai get same chorus, when if the parts before are at same structure, but as mentioned before, in a slow song 1 min maybe to short to get the whole chorus and as tryed, the ai is not able to detect a running chorus to fullfill it, when then part that is extended from has only half the chorus in it. Maybe some musicians see ai more of a tool to help to fullfill the creativity, but as today, it's more for the creativity of the ai. Hopefully it will improve in the future, so for producers this will be a tool to work for theire creative ideas, than giving up the creativity to the maschine,

[hybrid crane]

The one-minute limit on song extensions is influenced by technical constraints related to memory and GPU usage, and not merely an arbitrary decision. While extending this limit might seem like a straightforward solution, it involves complex development and optimization to balance system performance and creative flexibility effectively. If these current limitations are significantly impacting your creative process, it might be beneficial to revisit Suno once the software has more matured further and can better meet your needs. But the Suno team truly values your feedback as it plays a crucial role in guiding our development efforts. 🌼

[zinc hazel]

I know. My feedback is ment like that. By far it's the best plattform for such things right now, and I hope that it will improve more in the future.

[opaque kraken]

Hi, just discovered Suno.
Kind of wasted my first Points on trying to do a whole song, by extending something I already liked a lot.
I really miss an "Edit" option, where one could "edit" or "add" just a single word, or line of text/lyrics while rest stays mostly the same.

[tame swallow]

Like Fibo just said, it would be really nice to have an "Edit" option. Sometimes you come across what you're looking for, but come to find out, you made a typo, or you just need to adjust the lyrics to fit them into the given amount of syllables the AI made for a melody.

[opaque kraken]

I think I will fall in love with Suno, once it has all the small details needed to really create something. Right now it feels more like a demo. A really good demo, but still no really usefull.

[raw gale]

MAybe being able to switch part like doing an other part 2 but still using the part 3 created

[hybrid crane]

Creating a whole song doesn't cost you any credits 😉

[blissful condor]

Hello new to this kind of thing. i noticed in pregenerated lyrics it uses tags like [verse] and [chorus]
would the ai respond to things like [music break] [guitar solo] [bass drop] ?
if yes, please add it to the FAQ if no, please implement

[scenic kite]

Apparently not. I'm going crazy about extending a song by inserting a guitar solo or otherwise a completely instrumental part because it keeps inserting bits of vocals to fill in the base, despite using tags like [guitar solo] [instrumental only] [no vocals]. I don't know if I'm doing something wrong, but it really seems to ignore them.

[blissful condor]

between an edit feature and the addition of these tags. i think it would be a great upgrade. i'd pay for those features.

[scenic kite]

It is a pity. I am a premium user but on these things Udio seems a step ahead, although I generally like Suno's results better, at least the early generations.

[spiral acorn]

When they add "inpainting" you will get something like this. You will be able to select parts to re-generate.

[opaque kraken]

Maybe it can do the job for replacing just a word or 2 keeping everything else intact.

[errant jetty]

歌曲开头和结尾不是很好。对汉语和方言支持不好

[spiral acorn]

yea i don't see why it wouldn't work in those kind of situations. I assume you would be able to be very precise and select time regions like that

[spiral acorn]

There is a Chinese Chat here:
https://discord.com/channels/1069381916492562582/1154772025940922438/threads/1154780167047217233

Otherwise you need to use English for the other parts of the Discord

[scenic kite]

Hi, do you know if there is a roadmap somewhere?

[spiral acorn]

I don't believe there is one, at least not a public one

[scenic kite]

thank you

[rugged light]

I love Suno and would love to create several songs with my +3000 credits, however, I haven't had the availability, but I can't accumulate the credits for the next month 🥹 . This month, for example, my mother had an accident and I spend a lot of time in the hospital with her. It would be really good to be able to accumulate credits for premium members, at least half of the points would help, because I don't want to cancel the account 😕 .

[calm yarrow]

Hello where can i see the list of all the song i made ?because the history is only showing a few last ? Thanks

[hybrid crane]

Please don't double post.. thanks

[pallid ravine]

Hello all, I don't know if this is a feature or a bug, Signed up made a song about the Crusades, was flagged for moderation, made a song about Templars and had a song generated that mentions the Crusades, and Crusaders. Not trying to be super edgy but

https://tenor.com/view/princess-bride-using-words-dont-gif-22072801

[twin hinge]

I'd love to be able to add my own cover art. Also would be cool if you offered mixing/mastering inside of Suno.

[patent sonnet]

The ability to change the fricking song image.

[hybrid crane]

Someone asked yesterday and got an answer from a Dev. See: #📚┃suno-school message

[patent sonnet]

Ah cool, thanks.

[hybrid crane]

Why are you defaming this person? Even if you have a good reason, don't do that publicly. You can send me a DM about it. @pearl dirge

[trim wharf]

My songs are always 1.58 minutes long. How do I make more? When I browse the site, I see 3-minute songs.

[tepid trench]

Press the "extend" button. You can then generate up to another minute (well, 58 seconds to be precise) for the next part of the song. After this you can use "get whole song" on the last part to join them.

[trim wharf]

Thaankkss!!

[craggy echo]

I hope your mom will be OK. Prayers.

[spiral acorn]

You could try emailing support, maybe they can help you out with those +3000 credits? I'm not sure, but could be worth asking... even if doesn't work (it's maybe worth the try). Best wishes for your mom!

[remote rock]

I'll shoot my shot of suggesting these be thread channels so users can make suggestions/bug threads and not play chat roulette for their suggestion to be seen once again

Would also show which suggestions have the most activity/support

[patent matrix]

I have a suggestion.Can you be softer on moderation, because I wrote something with no cuss words or insults and it's still flaged for moderation.

[shut sparrow]

Not really a crazy suggested feature but can you make it so theirs is a repeat playlist button or auto repeat songs listening to in a playlist.

[timber totem]

This question sounds dumb but... how do you make a song popular ?

[kind cliff]

On the Suno website, it would be really nice if, when you're extending a song, you can see exactly what second in the song you extended from. So if you like 0:44 of a 1:00 clip, when you expand the "extended from" segment, it'll show the original clip and say "0:44" as well

which makes it easier to come back to a specific section if you start generating extentions on a different clip and then forget and have to relisten to find the exact moment (or if you're doing multiple extentions from, say, 0:42, 0:43, 0:44 and 0:45, and now you have no idea which generations came from which extention

[hazy fjord]

I really wish there was a HARDBASS in Suno. Like vodka, cheeki breeki u know.

[tacit echo]

I have 2 pretty simple QoL suggestions for extending songs 🙂

1. let us pick from any of the art generated from any of the clips when extending a song.

(a lot of the time when your trying to get that satisfying ending, the prompt you put in is very short and then the image generator can 'forget' the original lyrics and generates something completely unrelated.)

2. let us edit the listed lyrics upon song completion.

(the A.I. can sometimes misread lyrics or repeat things and it would be nice to be able to edit that in)

Sorry if these have been recommended already.

[merry quest]

Please deliver high quality downloads of 320Kbps

[solar vigil]

I would like to suggest myself as an intern. I think what you guys have made is epic and definitely want to be a part of it! @still parcel

[vital ginkgo]

Yes, Please let us Edit lyrics on Full songs!
Getting the generation to use the lines in V3 sometimes requires manipulation (like repeating lines) to get it to start where you want it to. This sucks in the Full song lyrics because its all over the place sometimes and definitely not accurate to what's actually there.

[severe glen]

How about letting premium users monetize from their playlists directly similiar to spotify but here on our own Suno accounts by adding a few ads to monetize? Also, ability for multi playlists...

[inner jetty]

How about letting the user give more specifics to the ai... such as male or female voice before a verse, that way you could create a duet, along that line... why not allow the user to define chords at specific points... a () around a chord in the lyrics would seem a logical way to do it... could also allow tempo specifics in brackets or something... As it is you have so little control over the direction the ai goes it becomes way too random.

[leaden flame]

Suggestion for Suno: Everyone would of course like to have better and more reliable user input control for Suno over different aspects of the music generation process; but, the question is, just how much user input control can be built into an AI music generation application like this? Just what is practically achievable and what isn't? It would be nice if a Suno developer would post a clear explanation of this on the Suno FAQ page. That way people could just be pointed to the FAQ article on this, instead of everyone asking for various user input control improvements here, over and over, every day, without it even being clear what is practically achievable. 🙂

[leaden flame]

Better yet, we should have the ability to download to a lossless audio file format such as .wav. 🙂

[tepid trench]

Better quality audio formats will come when the model produces higher quality output (one of the developers mentioned before that the limiting factor for quality at the moment is the model, not the MP3 format - they chose 192kbps because anything higher doesn't currently give any benefits.)

[leaden flame]

Makes sense. Thanks for the explanation!

[clever lodge]

I would love a few changes.

1: Folders in your library - I would love to be able to move my songs around to keep it clean
2: If a song is rejected for having something in it that is inappropriate, it should highlight that parts that need to change. it is hard when doing a rap to know what exactly is throwing off the AI from making it
3: Allowing us to edit how the song lyrics appear in a full song would be great, mine look really messy sometimes
4: a trial mode - I hate spending credits just testing if I can change the tempo, style or lyrics when making a song. Also to know what I can expect out of the song style before I create it. even if it's just a 10 second demo because some rock sounds amazing and other rock songs are almost impossible to listen to.
5: The ability after you make a song to then be able to make a Karaoke version of it. I want to sing along to my songs as well haha
6: the ability to select what instruments are used. It would be so cool to be able to use a tick box to select what instruments you want to use in your song.

[somber vault]

Pudiste resolver este inconveniente ? yo estaba haciendo una cancion para unos abuelos que fallecieron y me marca flagged for moderation 😦

[ebon flare]

I'm still a bit new but does anyone know if the order of what we put in the style matter? like if we put "reggae, rock", will it take the reggae as precendence over rock? Or does this have no impact?

[leaden flame]

It's AI, and from my observation there are no hard and fast rules for text input that work every time (at least currently). There is always a hit and miss aspect to this. The only thing people can do is experiment with different keywords and arrangements of words and phrases to see what is more likely to bring the kind of results they are looking for. If a person understands that there is always a degree of unpredictability to all of this with AI apps, then you will understand that it is always going to be somewhat of a trial and error process with this stuff. What works one time may not work the next few times you try it. 🙂

[wet adder]

It would be awesome to have a promotion channel where we could plug our SoundCloud, Spotify & YouTube links. Just purely based on promoting Suno music that is on other platforms.

[magic acorn]

Maybe to keep the quality of audio generations? Some of my song begin with a good sound and end like it was an AM radio station

[tepid bolt]

I would like the ability to add a few notes on the song- Like if the song needs editing I can say "edit from 3:52". This way I know what the next step is.

I would also like the ability to have a setting on the song that implies it is done- FINAL CUT; FINAL- could be very helpful for songs that are completely finished

Eventually I hope we will be able to edit the song we have created with more tools. I would love to be able to cut, add in more lyrics, add an ending or take two tracks and blend them together.

It would be AMAZING for the software to make a recommendation; Try making this song in the Style of Rock n Roll

The ability to make an Instrumental version of the song you just made

The ability to get a full list of the lyrics at the end of the Final song

The ability to slow a song down or speed it up;

a list of the Instruments, Styles, Beats, Tempo's, What Chord the song is in

Being able to adapt the vocals- I want a full on chorus, I want this part to echo, I want the vocalist to be a tenor and have a deep south country accent

[leaden marlin]

Make an Apple/Android App?

[reef wharf]

Hello i was request to post this here but adding the Erhu as an instrument. I've done of 20 creations and cannot for the life of me get it to come out.

[ebon flare]

Yes this! Plus also the ability to shuffle playlists

[leaden flame]

You can try a phrase like 'Chinese traditional double string instrument' or 'Chinese violin' and similar. Maybe that could work. Depends whether such an instrument or similar instrument is in the Suno model, however.

[winter wren]

I'm interested in using Suno to generate instrumental pieces that I can later edit in a music composition program like Sibelius or Musescore. So rather than an audio file, I'd like to get a MIDI file or similar out of Suno. Is this possible, I poked around and it didn't seem to be.

I looked around last Decemeber for AI transcription tools that will help me go from an audio file to MIDI, but the state of the art is not good.

Does anyone have a workflow for this?

[feral aspen]

Could try asking in channel #📚┃suno-school about Erhu

[leaden marlin]

For the next update their should be an mp3 uploader once u upload the mp3 the ai will think then create something similar to the mp3 or YouTube link

[granite lichen]

Add something where i can take the Beat from a other SOng into the new one with diffrent lyrics is this possible?

[leaden marlin]

Can i post my suno ai songs on Spotify?

@feral aspen

For free ofcourse

@tacit idol

[severe glen]

A search box to search growing library... and add hashtags for songs to be able to filter when searching

[mighty remnant]

It would be great to have a negative prompt to keep away things I want to avoid in a song.

[brisk cedar]

It would be great to be able to set a chord progression and/or BPM with time signature.

[tacit idol]

Are you a paid subscriber on Suno?
(Also, this post isn't a suggestion)

[crude zinc]

I think that it'd be nice to have a toggleable switch I can turn on to will allow me to pay the exact number of credits needed to finish all the lyrics in my song and have a proper denouement (or even have the denouement be toggleable on/off too). This way the songs don't get cut off two minutes in. I know there's the extend feature, but it feels a bit clunky to me. Also, 2 minutes just feels too short for a full song. 2 minutes isn't bad for testing the "feel" of a prompt, but I find it not quite long enough for most of the songs I want to make. I'd say mine would probably sit around 3-5 minutes.

If my idea does get implemented, it'd be nice to get an approximate on how long the song will be and maybe even some options for how to tweak the length of the song.

Also, a Thumbs Up and Thumbs Down (or maybe just a Thumbs Up button) in the Explore app could be nice for collecting data on what AI songs work well for training either your AI or the analytics of the client. What I mean by the "analytics of the client" is figuring out what the client likes to listen to and choose other similar music (I imagine this may already be somewhat implemented).

[tribal sentinel]

I really hate the way vocals are coming out. The first few lines may have a clear, solo vocal, but then the vocals are affected by nasty effects such as chorus or or multi-layering. So whenever I prompt for a [female solo vocal] that is what I want to get.

[rustic pagoda]

Hi, if our company wants to integrate Suno's API to generate music in real time, who should we contact for business cooperation?

[hybrid crane]

Yes.. but keep in mind, there is no api yet. You can send a mail to [e-mail]

[rustic pagoda]

Thanks

[tribal pulsar]

1. Mark Names
   We need a way to mark something as a name, so the bot doesn't use it as a way to change the type of the song. For example if someones name is "West" the bot will try to use it as a direction and not as a name.

2. Instead of instantly smashing out the creation please transcribe the song first and check if the lyrics are all in. It's really annoying if the song is REALLY good and the bot just cuts it of half way through or cuts off the last few words. Making us pay more credits eventhough it was not our fault

3. Don't be so overly strict on censoring. Rather turn off "Public" toggle if it detects "nsfw" stuff. Having a bad word in music is not bad. Let people create what they want in private.

[tepid trench]

Save the last settings used (e.g. custom mode, instrumental) and also don't reset things like instrumental toggle when changing between simple/custom. I have a habit of enabling instrumental mode before switching to custom mode and I have to then enable instrumental mode again. And I never use simple mode so remembering what I last used would be nice.

[wanton citrus]

Would be fun if the model could recognize music terms outside genres such as defined chord progression, some level of microtonality and rythmic figures

[oak cloak]

"Prompt not generated. Flagged for moderation."
Who came up with this warning? The IRS? xDD
Let me finish this prompt:
"We have used our advanced calculations to determine an exact matrix of acceptabilities. See if you can guess it on your next try!"

Using an AI for censorship is a great idea; not letting that AI highlight the offensive terms is oppressive and sucks up unnecessary calculations.

And the word was 'choke'. You know, there are a lot of chokes that exist in the world that have NOTHING to do with censored subjects.

[tribal sentinel]

Train the AI to create instrumental intro's when the [Intro] command is used.

[oak cloak]

You can start an instrumental by clicking 'instrumental'. You can extend it from any point you deem appropriate with lyrics.

[tribal sentinel]

I know, but if I extend that, it only gives 1 minute or less. So that is not the solution I'm looking for.

[hollow locust]

[Intro, Instrumental] usually works. If not, sometimes I spam [Instrumental] afterward. The worst offender is if you put 'female vocals' in the style box. Almost always starts the lyrics at 00:01 or sooner in that case.

I'd like it if it had more [] instructions. Things like [Emotional horn section], or [Sweeping strings with a throbbing bassline]

[slim zealot]

Ability to follow creators and get notified when they publish would be nice

[potent folio]

Allow creators to close downloads/continuation on their songs.

that small friction in downloading other's creations will cut down on the bot posters

[delicate cedar]

You may get a real kick out of #🐣┃suno-showcase

[quartz steppe]

just try it )
you have your prompt already - "hardbass vodka cheeki-breeki" should be sufficient ))
if don't get anything close in 100 attempts, try to request feeding data to the model
that is what I did for heavy dnb, and it worked )

[frosty venture]

Industry standard for a commercial song is approximately 3-5 minutes. While this tool can extend songs to get close to industry standards, I would like to be able to write a whole song and for the tool to sing a full 3-5 minutes without needing me to manually extend every minute and waste my time and money on constant reiterations of each individual extension.

[hazy fjord]

How can I make a request?

[quartz steppe]

well, I haven't found any way to do it properly, so, out of desperation, I simply reported one track as a bug, with this request inside
and very soon I begun to get some results, that are clearly based on the right data
I had examples of crossbreed, techno dnb, darkstep, hardstep, also one instance of late 90's techstep

[sinful valley]

A few things: I am not sure why but nearly all my v3 instrumentals have ghost singing and its not useable. Just my opinion but i dont think V3 is ready, the success rate is about 1 in a 100. I put a lot of effort into the structure, the lyrics expecially, especially when that is the only control point i have. Yeah if someone wants to just muck about then the system may suit their needs, however i utilise it to help seed ideas for me to work off when i am blocked artistically. Harmonies etc. The screams of the underworld really puts me off.

I understand that AI is i suppose in its infancy, however at this point i would have thought that there would be the ability to feed it some info. But when meta tags or even just song structure markers tags are inaccurate or dont work, then she got a long way to go.

It was a cool trick, but unless it supports creativity instead of just mimicing echoes of it, it will eventually be the tamagotchi, people will say, remember when we thought we were cool, for five seconds.

Not only would i like it to work as its generally advertised , i would love if i can input a structure, approx time codes, notes, instruments. Surely the fact we are doing half the work will ensure something less hap hazard is produced.

Lastly i would love to know why it hallucinates? Why doesnt it perform some sort of check back processing and reference material relative to music theory, the internet has everything, or is it quarantined to avoid it becoming skynet singing at us to death.

yeah for the whole library

[slim zealot]

THIS

[quartz steppe]

for this reason, I hope there would be a possibility to refine tracks in some future versions, where model could try to improve quality of what was already generated

[sinful grove]

Suggestion 1: Time of the new segments should be able to be on a slider that increases the usage of credits as you increase the time of the extension you want. This way you can make a longer segment or shorter one if you need something but also Suno can then keep the credit system the way they currently have it.

Suggestion 2: It would be cool to have an option other than "extend" that would allow us to keep an extension or segment but maybe give some feedback and the AI can take that into consideration to recreate the same extension. For instance, maybe I really like an extension but I want the chorus to sound the same as the previous segment. It would be great to be able to tell the AI that I want the Chorus to sound the same as the other one and then regenerate the same extension (I would still be okay spending the same credits on this as if I were making a new extension)

[sinful valley]

i think the issue is it doesn't know what to improve because it thought it was giving you music in the first place. "you mean you didnt want athereal organ piano harpsichord drum trumpet as the melody?"

thanks guys x

[chilly kelp]

If you find some issue with the lyrics, there is currently no way to create the song to fit your needs. ''Reuse prompt'' does not generate a new version of the same song, it creates a whole new song.
I would love in part 1 of 3 to edit the text a bit, but currently there is no way to change that. A ''Edit lyrics'' or ''recreate song'' would be a great addition, when the AI uses the already generated song as a base, and simply takes the new written text into consideration for the new version, so it'll be the same song but with different lyrics or setup

[wooden fable]

Try [Instrumental Intro] ... it seems to work for me

[grizzled pine]

Hey everyone, first of all, Suno, well done 🙂 I enjoy the music generated! I wonder if it would be a good addition to be able to say you don't want a beat or drums in the generation. Kind of a switch like the Instrumental switch. I feel that when I want to generate ambient music, there's always a beat coming whatever the prompt is and it would be great to have a parameter over that.

[wooden fable]

Yes, indeedn. That's why I think a "negative prompt" like in Stable diffusion would be cool: So that you can tell the AI what you don't want, e.g. "drums" or "autotune" or such.

Another thing: Does anyone know what's on the future roadmap of Suno? What are the devs gonna do next? As a paying user I feel it would be really nice to have a bit more transparency, and hear just a little more info relayed from the devs. Just to get an idea where this is all headed.

Suno feels a bit like a black box tbh.

Basically, more #🚀┃updates please. 🙏 Let us know what's going on.

In particular: is the choir effect (fuzzy reverby voices) on the roadmap?

[vagrant juniper]

When making an instrumental, it would be nice if there was a way to end the song. Like add an ''outro'' button like Udio has. Nevermind, it seems adding (outro) in the ''style'' prompts does the trick.

[unkempt geode]

Need paypal for subscribtion

[vapid isle]

please add code to " restart" Song, both vocal and instrument. why ? i/ we have made a Song that became epic vocal and instrument. but i accidently had wrong lyrics,or i changed it later because of better meaning, i would love to start this exact same song ( vocal, instrument) but just with new lyrics. no lyrics are final, they change a lot and its difficult to find best when your song is changed every time. it would help sooooo much - just suggestion, no hate pls and yes i asked this one time before. be patient 🙂

[vapid isle]

there is a [end] [fade out] [big end] command, take a look at suno wikki

[vagrant juniper]

Thanks! I'll check it out! 😉

[main mango]

This. After a while I like to evolve the lyrics to better suit. But if I make a typo I can’t fix that

[vagrant juniper]

Have you tried this for instrumental's though?

[night jackal]

Can’t u just make an extension?

[still stirrup]

oh my god, people really have to beg for things like this???

it should be so basic

[hybrid crane]

It's a young company with a small team and still developing. The have to set priorities.. Every software company have to start somewhere..

[still stirrup]

shit like this require like 10 mins to add or something

[hybrid crane]

As always.. "The best helmsmen are on shore".. 😁

[wooden fable]

i wonder what the devs are working on atm.

i'm sure they're busy

but it would be interesting to know what's on the top of their list

and/or to know about their plans for the future

just a few words

[still stirrup]

nvm, udio is better anyways

[tiny sleet]

It would be cool if it is possible to get the same voice for multiple generations.

[hybrid hinge]

I'm really impressed at how Suno has evolved, but one feature I would like is to be able to follow other profiles. There is a lot of great music on the site!

[still stirrup]

you cant even search for songs or genres

the home page is limited to just a few songs

you cant download other peoples creations

ok im done hating i got bored xd

[tepid bolt]

Have you seen this? I'm thinking about using it... Don't know anything about it, but it seems this is how independent artists get monetized on streaming/app services. https://www.tunecore.com/sell-your-music-online

On that note- Suno, to keep this easy... https://www.tunecore.com/sell-your-music-online, I would rather stick with your platform to publish. You have the ultimate bundle available to you; Keep your users right here on platform navigating from idea to monetization.

[grizzled phoenix]

Can we have a download songs feature?

[tepid trench]

You can download your own songs

[grizzled phoenix]

I mean others' songs.

IK I can download mine

[tepid trench]

I was just about to edit to mention that (sorry!) - this used to be possible but it was removed as people were basically stealing the trending songs and uploading them elsewhere as their own.

But, I do miss being able to download other songs, even though the playlist feature seems to be the intended alternative.

[grizzled phoenix]

what if I don't have access to internet 24/7 and use a phone that's older than me?

which can't handle the internet

[grizzled phoenix]

Maybe the devs can add something that checks for plagiarism

IDK may not be so straightforward.

there's too many variables

length

title

prompt used

how old it is ig

but yeah if done, it can be done

[wooden fable]

oh please don't ask for even more censoring and filters.

[tribal sentinel]

I've tried, but it doesn't work for me

[wooden fable]

oh? hmm. well, nothing is guaranteed to work. this is a casino after all.

i just had luck with it one some occasions

just like, i've also chosen the "instrumental" option, and still gotten a choir.

AI bots – they're like herding cats!

[grizzled phoenix]

nah not censoring

I mean if copy pasting is all people do, it's better to take control of just that one thing.

[leaden marlin]

the ability to get a snippet and extend a song would be cool

[minor portal]

I would love to be able to sing my lyrics and get an accompaniment in a chosen style. Just having my lyrics with any melody is fun, but I'd rather be able to also have exactly the melody Ihave in mind #-)

[wind abyss]

a separated parts page from the main library, eg; like an work area, so created parts end up separated from the main library.

that way you can keep parts you want in the future.

[lyric idol]

Would be nice to be able to edit lyrics on full songs so you could remove overlaps from stitching

[frosty venture]

Please add a DELETE ALL PERMANENTLY option to the trash bin. I do so many iterations of a song just to get it how I like it, the housecleaning alone to delete the excess trash is a job in itself.

[leaden marlin]

No i don't want to monetize my songs

[frosty venture]

Two days of making songs and I had neary 10 pages of songs to delete

[tepid trench]

I often see people ask why they can't create a song, and it turns out it's because they have no credits left.

Instead of simply disabling the "Create" button, either replace it or relabel it to say "You do not have enough credits", "You need at least 10 credits" (or similar)

[eternal hatch]

While the AI is an AI and not easily tamed I really think improving the interface is one of the simplest and cheapest ways to improve the service. It is even more important for sound than for text or images - it is so much easier to follow a thread of thought or scan your images. But trying to remember which version of a song is which is soon impossible. Structure, folders, tags, sorting, filters. (And v4 🙂

[summer vapor]

Does anyone know how to always get the same voice on different songs?

[minor portal]

You can extend a song you like from 00:01 then you get the same sounds again for the new entry, but only 60 seconds.

A needed feature would also be to merge the extentions to the songs, otherwise you can only publish songs with cropped ends.

[pallid ravine]

Nadie me ha respondido, por eso publiqué. No sé la gravedad de la bandera automática. ¿Si son tres strikes estás fuera? o simplemente no uses estas palabras. Me gustaría arriesgarme a hacer canciones mientras busco minas en un campo minado.

[tardy wolf]

Make an 18+ version without this pg13 bullsh... God dammit.

[solar cradle]

Allow us to export songs with seperate midi and vocal tracks to be able to import into music software and keep working on the song.

[inner junco]

para para para iki yüzüde kara

[hybrid crane]

Uhm?

[craggy echo]

Not sure if your ? is because of what he said or because you don't know what he said. If it's the latter, it's "money, money, money on two sides" according to my browser translation.

[hybrid crane]

Yes I translated it.. but still have no clue why it is posted in Suggestions

[craggy echo]

Maybe we need another sub-forum in the Feedback section: Complaints. 🫤

[hidden ibex]

Instead of like or dislike there should be a 5 star rating.

[wind crest]

would love if flagged content was hilighted. i've written a song that is suggestive within normal radio standards but not explicit, after it got flagged i pruned a swear word, after it got flagged again i cut or toned down some other phrases, but it's still getting flagged. instead of just cutting bits randomly it'd be extremely convenient if the problematic portion was noted in some way so it can be changed to something acceptable without butchering the entire thing at random and getting flagged a million times

[lean island]

Based on own research, exiperiments some new feature wishes 😉

• Control over normalization strategy (if applicable) (incl. db headroom)
• If possible: continue (from to) - so we can keep a good part that comes later, replacing bad stuff before

[inner coral]

Indeed, I've suggested the same feature several weeks ago and was met with vehement criticism from a few individuals who appeared to lack cognitive depth. They argued that it would overly complicate matters and deemed it unnecessary... such attitudes are quite disheartening.

[vapid isle]

just select all and go to the ... 3 dots and select "to trash", after this you need to delete them from your library as well, voilá you are happy 😄

[vapid isle]

its works at any kind of music, maybe you need some generations. sorry but in my case it works very well

[vapid isle]

nope, extension need a time to set, but 0:00 dont work, and 0:01 cuts the first letter so its corrupted beginning every time. you are not able to start a song from scratch using same vocal and instrument arrangement. thats why i need this. at this moment you cannot make music in professional way, you need to gamble it and tthats why m stuck, i need to tweak all the things i need to finish my song in a professional way to share it on high level platfrom. yeah thanks buddy

be sure its PEG 13, so you have to soften your words

[spiral summit]

V4

[delicate cedar]

Yeah, this is a violation of our rules so obviously this was never going to be granted.

Rule #1️⃣ and #2️⃣.

Since you're new here I'm going to encourage you to reread the rules and ponder why we might consider a song like that to be bullying or hate speech (never mind what Discord itself might think of it) and go right on considering that for a while....

[dawn fjord]

Bulk. Download. Please.

[iron hare]

For future suggestions I'd like to add one: What about a function that you upload the instrumental mp3 and the IA generate the singing too? It would be a hella of help for producers too

[ruby dagger]

Would love to be able to search for stuff on the site. Specifically, search for 1) a genre category, and 2) a title of any piece made public. I can't see a way to do either currently.