💡┃suggestions-web

[honest bridge]

Website related suggestion: There doesn’t seem to be a way to close the banners on mobile (e.g. there’s one up now about subscribers only during updates). It would be nice to be able to close that as a subscriber, heh

[lunar marsh]

It would be awesome to be able to regenerate clips using the same seed but with modifications—like, same vocals and cadence, different lyrics, as an example. Sometimes the music is perfect but the vocals are wrong and it’s such a bummer to then have to regenerate another ~20 iterations to get something that ends up working in a similar way

[untold shuttle]

Trying to work with instrumentals and it's frustrating that I can't find a metatag that stops the melody and only has the background music presented. a metatag would be great for that work. Have tried intro but if the melody starts immediately then it’s no way to stop the melody to playing.

[turbid ocean]

Are they flagging your music when you upload it? You might need to try cleaning it up and exporting it as a WAV because I'm guessing it's flagging the quality

[reef marten]

a couple more suggestions:
acapello and karaoke exports
right at the moment i have vocals I want to use, but i want to remix the backbeat i ableton
but sometimes there's a lyrical issue and i'd rather just sing over the flaw, or generate new vocals to cover it
I'm sure a lot of other users would have no end of uses for these two features

[harsh cypress]

The trending page need to have time based sections like "Today, this week, this month, all the time", and needs to have a next page button

[hollow locust]

it drops things like [intro, instrumental] and starts the lyrics at 00:00 coz I think it's trying to cram the whole song into 2:00... Does it realize it doesn't have to? Then when you 'continue from this song' it seems to assume that's going to be the end, and sometimes tacks on no more than 00:10 of song time.

So... I guess my suggestion is... have an option/switch to tell it if the continuation is ended to finalize the song or not? Or something like that.

[hybrid dragon]

You can use the gender in the "Style Of Music" prompt

[nimble trench]

Request: Please provide a pay-per-credit option. I would be so happy and excited to recommend Suno to everyone I know, but having to sign up for a monthly subscription just to receive credits that expire and don't rollover is a really scummy business practice. Utility billing is the appropriate way to charge for something like Suno, and there should be no monthly fees. With a change to a pay-per-credit model I would tell everyone to use Suno. At the moment, I advise against it because monthly subscription billing for expiring non-rollover credits is just too scammy for me to recommend to anyone I know.

[spiral vessel]

Revice a Song by Instructions

To Suno:
please find attached a suggestion on how the UI should be improved and why there should be a new section called Instructions.

To the users:
If you like my suggestion, please comment with suggestions on what instructions you think should be there and what you have been missing so far
Thank You!

[astral copper]

It's not the quality problem, I have both mp3 and wav files. Also most platforms are fine distributing the content. It's more about trying to claiming the content on the platform so that I can add lyrics to the songs where the problem rises due to each platform's policy. In this case, Tencent requires more proof that I own the rights to the songs, which since the right is granted by Suno to me as a paid user, I don't have any proof to show I own the song.

[hollow locust]

But what if you want a female solo for the versus, and a male chorus for the actual chorus?

[scarlet sequoia]

Specific Style of Music Customization Controls.

Prompt is good but still would be nice.

[still holly]

I got a suggestion, can we upload a custom voice? i think it would be neat if we could make markiplier sing or even make myself sing well

[dusty raft]

👍

[summer hollow]

you should add the option to add vocals on a instrumental whe made so whe can upload the instrumental and add some vocals on it

[untold shuttle]

Something that have to be improved is to use the same melody and chords when continue. It’s incredibly annoying when you have a perfect groove that you will use during the whole track and it changes every time you continue. It doesn’t happen all the time but too often.

[summer hollow]

maybe add the function to give the ai a song so the ai know what kind of song whe want if it has lyrics/ voice the ai can also know what kind of voice whe want @hybrid crane

[waxen creek]

Suggestion:

Can the default white text on a black background be changed/reversed? It gives me a migraine after a while.

Many thanks.

[vestal sinew]

it would be really really great if it highlighted or just said which particular lines or words have gotten my lyrics flagged for moderation.. it literally took my 10+ tries of deleting or changing words to find out the offender that caused a moderator flag.. and i believe the phrase that did it was "Soaking wet"... it would really save me time and possibly give me a chuckle about which phrases the ai thinks may be offensive or against policies ..

[hallow niche]

Intergration of other paymentes( IDEAL, PayPal )

[lyric heron]

Suggestion:

let us see a list of the songs we liked, i can like songs but i can't find them later, i can't see a list of the songs i have liked before
you can do something like Youtube does which is a private playlist called "Liked" and any video that you like gets added to that playlist automatically (or Youtube's WatchLater feature which is just a Playlist)

Why :
as Explore tab is bad and not really useful right now, if i want to find a song that liked, i can't really do it if i don't already have a link to it or have not manually added it to a playlist myself

alternative approach: make Explore tab so much better but i think that requires way more features being added and is something that might happen eventually but i think adding a way for the users to see what songs they liked is much more easier to implement

[swift jetty]

i would love a new feature where we can upload our voice to use in the songs.
And also upload a song style for example in Curacao Caribbean there is a genre called comeback( its like a type of cuban montuno song)
But singed in language papiamentu.
would be nice if we can add 3 songs of that genre than it learns or something or use similar beat to make our songs.

[tame flume]

Hello Everyone,
I enjoy suno music generation so much. I would love to see the website be more social, like we could follow each other, having access of their published music, like them and leave a comment, the user would have a feed-back of how many likes/comments received and based on likes / music published, etc., the user would receive some free credits. The public music would be browseable and filterable. Thank you for your consideration 🙂

[narrow talon]

You can already do this. Click the three dots behind the Song and ad it to the Playlist.

[unkempt cedar]

Zuno is amazing, I wish the developers would work on the "continuation" feature or even better extend the original time for the song it creates. As it is the continuation really breaks the flow and no matter how you do it you then need to get both files and try to merge them ( i am a noob) but thats my suggestion...

[feral aspen]

On the continued clip (or the final continued clip) if you click "..." you can select "Get Whole Song" and it will merge the clips for you. (Does not cost credits.)

[unkempt cedar]

Noobness confirmed!!!!

THanks

[wary swan]

Hi bro! Possible u can tell me how to make an emphasis via this net? How i can highlite needed words / acents

[gusty silo]

Could you add a [Hallucinate] tag for when we actually want it and at the same time don't allow it to do so when we continue with our own lyrics input. Also please update the FAQ, and add all tags, all genera, and all instruments it can create (it doesn't understand didgeridoo for instance). It far to hit and miss atm. Stabbing in the dark, which admittedly does come out with some really cools things sometimes... Some additions to existing tags like [Intro 10 seconds] [Voice 1 male] [Voice 2 female], sure you get the idea.

[feral aspen]

A fellow hallucination appreciator 👍 . One thing that can work partially for your first suggestion is the tag [Incomprehensible]
https://app.suno.ai/song/62719a69-7623-4289-aae5-e776e8a603d6/

[gusty silo]

[Incomprehensible] lol didnt know about that one. we really do need a full list of known tags

Example of it just making it up after i gave it lyrics for the continue https://app.suno.ai/song/b152f236-a2e3-4fb2-a532-324c209b5d43

If you want really random lyrics https://app.suno.ai/song/c9872d76-71c8-4598-81d8-5952922b8aa9

[opaque sapphire]

It would be great it we could upload a specific song (even a previous gen) to make sure the next gen is in the same style. That way we could make soundtracks and albums

[twin swift]

• set bpm
• set key
• set key change 👉 [Chorus one tone up]

[opaque sapphire]

Also, control over how closely/loosely it follows the lyrics and structure would be great

[gleaming beacon]

Shared the app with collegues this morning and sadly "generations are only enabled for subscribers". missed opportunity.

if people go on your site and they can't generate music, they will never come back, so consider anybody visiting the app for the first time while it's in "sub only" mode , a wasted opportunity.

Suggestion : invest in hardware so you are never overloaded. I am sure it will pay off because your tech is amazing and everybody loves music.

[tepid trench]

It's a nice idea but there's already people who game the system to have more free songs. They'd just end up with people clearing their browser cache/storage repeatedly and forever being a "first time visitor".

[mossy oriole]

A folder system for songs would be helpful to clean up and designate song files to a folder, for potential future use.

[west iron]

When generating a song based on the prompt, I'd love a way to see a preview of the title and text, and option to edit, before submitting it for audio generation. Right now I just use chatgpt for the text generation instead which is a shame

[gleaming beacon]

Valid point, you could deal with this issue by checking for unique IP addresses.

I tried to tool last night for the first time, and I wanted to buy a submission this morning, but I saw that servers were overloaded and I was unable to show this amazing tool to my collegues.

Moreover someone on discord said they are a sub, but still unable to use the tool.

For this reason I decided to keep my money.

just my 2 cents.

I think you guys have the potential to become absolutely viral, if you make the right choices.

[tepid trench]

They already seem to have gone viral (people are talking about it where I work, friends are asking me if I've heard of Suno), which is probably why their servers are under so much load recently. Free accounts get blocked temporarily so that subscribers can use the service (rather than nobody be able to.)

[spice glen]

give a option to make only one generation instead of 2 so it gives the users more songs

also add a premium and premiere chat

[potent rain]

Hard +1. Id like to chat with other power users who are generating dozens or hundreds of songs per day about techniques and what works without the barrage of help requests and music submissions (that box out all other content)

[feral aspen]

yeah the Chirp-Technical channel is not used very often, and is meant for that. (will try to get it renamed to music-prompting-technical or something more obvious) https://discord.com/channels/1069381916492562582/1102676217687195679

[tepid trench]

I always assumed bark and chirp technical were meant to be about the implementation details (later realising chirp isn't open source anyway)

[potent rain]

Same

I think we could use a postproduction channel too, so people can discuss the stuff people do afterward like stemming and getting it into a DAW

[feral aspen]

discord is due for a reorg eventually, for now just working with that we got

[potent rain]

What do you use to get it into a daw and edit? (I’m a music noob so trying to figure that part out)

[mystic arrow]

Hey Suno Ai team I think you guys found product market fit. Everyday I wake up I want to use Suno AI intensively. The UI is awsome too straight to the point and more importantly, after some trial with some others Music Generation AI your output finished product is top tier.

[potent rain]

Sounds like I need to pick a DAW first

[golden hull]

Would be awesome if the exported video had a karaoke style text.. just so that you can follow the lyric easier. Right now, without anything it’s kind of hard when the lyric itself isn’t that clear.

[glass portal]

Suno Team: Any timeline on when we can use Suno via API? Would love to start building some products using Suno as the engine

[queen lichen]

Lifetime offer (LTD) to purchase a special plan ❤️

Pagination in Explore tab to find more musics and inspirations

[potent rain]

Also +1. Im making my own UI and the api backing the UI is simple enough to easily reverse engineer but itd be nice to have an official one.

And not have to play games with CORS

[visual brook]

for multi generation songs, it would be really nice if the UX/GUI showed each time a new gen started (timestamps), and the prompts used for each segment

[glass portal]

Agreed. CORS is solvable by routing from the backend, but would be nicer to have official

[kind crow]

easier way to delete multiple of bad generation at once, also I'd like a better generation of touhou style music, it seems like the AI recognize what it is, but fail to make a style similar to it.

[late plaza]

1. Create one instrument loops.
2. Upload midi-files and tell it to generate a guitar for example.
3. Or maybe just Audio-to-Audio, but that would not be as precise.

[merry hamlet]

It would be great if a user could upload their voice model.

[old minnow]

adding functionality for users to follow accounts rather than just a trending songs. For example, I follow x user so that I know when they post a new public song.

Also wish there was some sort of better way to find music. Like trending per category or trending with this tag or something. It feels like once a song moves off trending or new they effectively no longer exist

[lament kite]

Guys, do you have API?

[gray musk]

The only reason i don't pay the sub yet its because there is no option to do it with Paypal. it will be better if you can add the option and its more secure to the users

[queen brook]

I wanna search songs via genre and add to my playlist

[spice glen]

can you post the music genererated

on spotify

and platforms

if i bought subscription

[lethal cypress]

Yes

[little copper]

...

[lethal cypress]

Would love a feature of reference track so it can train on experimental Genres and more on song structures

[wanton thistle]

1. 🎼 Key
2. ⏳️Time signature
3. 💓BPM

PRETTY PLEASE, WITH SUGAR 🍬 AND HONEY 🍯 AND WHIPPED CREAM 🍦 ALL OVER, AND CHERRY 🍒 ON TOP! 😭😭😭😭😭

Stems would be great, too, but we can split for now on Bandlab and such

[kindred cargo]

When will this thing pass?

[north dune]

Can we have the V3 Alpha option? After burning through over 10k credits since V3, I've had better results with Alpha. I'm not the only one.

[feral aspen]

When people suggest "user uploaded voice models", does 'model' just mean voice audio sample, or does it refer to an outside AI model like RVC? (Also if this feature did exist, I think the coolest use would be uploading non voices and seeing what music Suno comes up with when you ask it to sing with the sound of a thunderstorm instead of a voice.)

[potent rain]

audio sample. most people using suno probably have no idea what model weights even are

and agreed. i'd love to sample some of Hans Zimmer's new stuff from dune and see what would happen

like the Harkonnen speach

[feral aspen]

I burned many credits trying to get Suno do something like that, just with prompts, by interleaving lyric word [instrument name] lyric word [instrument name] and trying to get a lucky gen where sounds like it mixes them up

[potent rain]

what was your best result?

[kindred cargo]

"no one."?

[feral aspen]

I have too many songs, hard to find... this was the first one that I thought sort of worked (v1), the first line specifically.

[feral aspen]

singer holds notes super long like bagpipes maybe https://app.suno.ai/song/23bb8d5c-c0d5-44ce-9ba7-7d31fe25a3c0/

[tight beacon]

"Permanently Deleted" songs should refund the 10 credits. Not sure how they could work that in though. But it would be nice to get those 10 credits wasted on a bad generation you didn't like because it came out wrong. Sadly would probably open the doors to abuse.

[sage mirage]

the problem with that is the GPU has been used

[wintry kernel]

more payment options like paypal for people that dont have a credit card.

[sage mirage]

Hitting 'generate' rolls the dice, slurps down too many electrons, generates heat and makes a GPU go "wowowow" and you get a result

Financially it makes no sense, especially since you do get the audio

[tight beacon]

That's not what the problem is. The problem is it's hard to tell if a song deletion was because of just bad generation the user doesn't like and someone just exploiting it to get free credits. I.E. downloading a song then permanently deleting the song to get the credits back and generating another full song.

[mental barn]

Usually, a small little sample may not sound so great, but RVC still needs a drive voice to get it singing. I thought about getting crowd noise and trying that, but once in awhile, Suno will make parts that sort of simulate that kind of thing, but it is quite difficult to control.

[warped finch]

Probably something where you can upload your song and let suno ai edit lyrics or idk

[turbid ocean]

@formal bear Hi Tony I saw your latest poll regarding how the 'Like' button is used. I think the options presented are incomplete and may skew the results. There should be an option for "I thumbs up tracks that may not be high quality but I want to come back to later"

[mild pier]

Hi everyone, New here. Working on a song and almost finished it but last part (3) won't complete, it just goes to a quick beat and stops,.. rest of song not done. I put in prompts: break, guitar solo, chorus, big finish. Suggestions please?

[hybrid crane]

MelRock99, it seems like you're experiencing an issue with song generation on Suno. Try refining your prompts with specific metatags like [End], [Fade Out], or [Big Finish] to guide the AI towards a complete song ending.

[potent rain]

when you say prompt, do you mean the style of music prompt or the lyrics prompt

[potent rain]

try continuing from earlier in the previous song

[mild pier]

I did that. I used teh [ ] around the words. My first two sections were fine, just this last part didn't respond.

[mild pier]

I cont'd from part 2. So I have to con't from part 1, correct?

[potent rain]

no, i mean try continuing from earlier in part 2, perhaps where there's more activity. if you continue from a part that's silent or fading out it thinks its the end of the song

[hybrid crane]

@mild pier, you might want to try leaving both the lyrics and style of music fields empty, and simply type end in the style of music field. Remember, sometimes it takes a few attempts for the desired result. In some things there are more options.. if [metatags] do not work, this might.

[hybrid crane]

We call the "style of music" + "lyrics" the prompts, because everything has influence.. Even the lyric itself.

[terse pendant]

Feature request/suggestion:
Playlist shuffle
Playlist repeat
See number of plays in your library and on the song detail page

[harsh fossil]

Feature Request / Suggestion:
Solo Vocals (Opposite of the "Instrumental" toggle)

[lyric heron]

you didn't understand what i was asking for
I know that we can manually add them to a playlist, but im saying it would be better for the User if they didn't have to manually add every single song they like to a playlist, you can see full list of all the things you've liked on any social media (or any good functioning website that has a "like" feature)

what is the point of liking things if :

1. i can't see a list of the thigns i've liked
2. you cant see what is the most popular or most liked song (Explore tab or Search)
3. there is no search function (again Explore tab)

basically i'm saying there should be some functionality added to the website to allow users Explore better and find other songs more easily, but thats something that i think they have on their plans and it takes time to Develop a good Explore tab with good Search functionalities, so why not for the time being atleast allow users to see a list of their own Liked songs

[tall glacier]

How can I use the same instrumental but update the lyrics? Anyone knows? Also, why am I getting Couldn't generate that.
Prompt flagged for moderation

[harsh fossil]

You can see a list of your own Liked songs.
Library > [ Liked ] , the button is right there.

That being said I agree there should be more fine tune controlled for searching through your own songs.
For example you can't search by name, which is silly.
You also can't search by style tags, which would be very helpful.
It would also be nice if the root song was at a higher level than all tracks continued from it. Makes it easier to resume a project that you're piecing together.

[lyric heron]

Suno AI is cool, but maybe i don't want to use it (or i'm out of Credit and Ideas) so i just want to see some cool things other people have done with it, which is why Explore tab and Liking other songs exists
but if my likes don't mean anything, even for myself (i can't see what i've liked), then what is the point of having like feature ?

[lyric heron]

no you can't, Library > [ Liked ] only shows you Your Own songs that you have liked, not songs that other people created and you've liked

it's basically a way for you to filter songs you've created

I was talking about songs other people created and i liked them because i saw it on Explore tab

and i know its easy to implement this feature which is why i requested / suggested it
they can easily get "your liked songs" from the database or many other ways like adding them to a private playlist and then show that playlist to you

[heady sluice]

song interpreter ?

[opaque gull]

Here are my suggestions what could be useful and added: an entry field where user can enter the desired length, a field where the desired arrangement of verse, chorus, bridge, brake, guitar solo, etc. can be entered, entry and modification of the desired voice, using one's own voice, deriving styles from previous created or uploaded songs.

Choice of instruments (also for the different song parts)

more influence on the outcome

[fickle valley]

Can we have a contest of AI by genre?

[rugged pendant]

I was on the verge of subscribing, but not being able to set the tempo is a deal-breaker for me. Including the key in the style (in Custom mode) seems to work most of the time for instrumental only but not for tracks with vocals, based on the tests I've done so far. BPM seems to be ignored though.

[gleaming beacon]

One one hand servers are so overloaded that you need to disable the AI for non subscribers, potentially losing tons of potential customers that can't demo your product. On the other hand you guys allow "10 running jobs at once".

What's the point of generating music faster than you can even listen to the damn song, especially if your servers are overloaded

[potent rain]

The "Due to system upgrades, generations are temporarily only enabled for subscribers." bar should be hidden for Pro/Premier users or at least have an x so we can close it. it makes the app unusable on iphone

[opaque gull]

Option if fade in and/or fade out is wanted or abrupt beginnings and endings.

[short reef]

I wonder.

[lethal cypress]

Yeah

[lethal cypress]

Yeah Would love that

[leaden marlin]

Add a “Close the song” button that will make ChatGPT make closing lyrics for the song

[feral aspen]

In custom mode this usually works if you leave your last line of lyrics and then write something like this in the lyrics field above it:

Write the final closing verse of the song, that comes after this verse:

[Verse]
blah
blah
blah
blah

[leaden marlin]

Okay thank you!

[random sparrow]

When looking at the new song in Explore, it would be useful to have a search. So we could select by style, language, etc...

[mellow tide]

Suno team, not sure where you guys are taking this long term, but this might be an interesting crossover:
https://www.youtube.com/watch?v=bnCR-SVB254

[fallen forum]

When merging songs lyrics often get messed up, add the ability to edit lyrics for "whole songs" before video download so that the lyrics in the video follow the song

[safe flare]

Yeah I literally just thought of this as a suggestion... but you beat me to it... A way to edit lyrics would be great as the merges really mess with things.

Although if I scroll up far enough I'm sure we are not unique...

[fallen forum]

Another suggestion would be to make the songs actually 2 minutes when the text is that long. I often give a long text and I often get 40-70 seconds.

[fallen forum]

Most likely not. I did not search

[opaque gull]

Optimize/harmonize function of existing created songs. (sometimes the voice changes throughout a song)

[sick meteor]

Suno could give the users who sub chance to buy our "free" generated songs. (Please 🥹 )

Maybe a feature what allowing us to treat an engine on our own voice so after inputing the lyrics we get a generated song with our own voice would be crazzzzyyyy : )

now yearly I do one non-commercial music album but if the above would happen I would be like Taylor Swift xD

[vague oak]

Am i the only one who can not contral to revise a son? Am i wrong? Is it all possible?

[wheat sorrel]

I would like "Saved prompts and lyrics" sometimes ppl just do other stuff, having stuff still there once you come back, Would speed up the process a lot, Like i wanna go sleep rn, but i dont wanna bother saving Prompts Etc in a Word file just to have em all back tmmrw, Also maybe a Note Field so people can make Notes of what Prompts work for them and what they do etc

[wheat sorrel]

This would be so cool, i Imagine you would have to Record a Lot of Specific audio and it would barely work, but i love the idea

[unique surge]

some prompts can give vastly different sounding results, would be nice if you could tell the AI upon remixing to be "somewhat similar" or "very similar" to the original track

[elfin echo]

Quick suggestions.

1# You should add milliseconds time stamp for continue from this song and milliseconds counter to songs.

2# It would be great if you can select parts of the generated song that you want to change. Lets say you had great flow on your generated song, but ai didnt pronounce some words correctly or you just want to change some text

[bleak summit]

Pleeeeease! Add option to exclude instruments. Every time when I get saxophone I have need to pull out all my hair! 😭

[elfin echo]

You can try to add it in your [] tags. For example [Instrumental without Saxophone] or [Instrumental, Exclude saxophone]

brackets pretty much work like mini promts inside your lyrics and you can influence ai song generation thru that

Might be unstable, but should work

[bleak summit]

I tried in Style of Music, "exclude saxophone", then i got it more often. I try also "[Chorus: Solo]", at least I won't get like 100 voices

[elfin echo]

style of music is buggy as hell

out of 10 times i set style as reggae, I get reggae instrumental, but trap vocals

[bleak summit]

yes, I ask for Soul and sometime I will get Rock 😕

[elfin echo]

So I specify it in [Reggae Vocals, Jamaican accent, Hook]

[bleak summit]

the downside is less characters for lyrics ... I think I will stop for today, 1:44AM... time to sleep. Bye and thanks!

[elfin echo]

yeah, but you only need to specify it once on first hook or instrumental

[neat gorge]

Can we please please please please please get some damn stems already.

[neat gorge]

Stems would fix this, lol

[elfin echo]

Its really easy with new fl studio

[tepid trench]

As I understand it, the AI is trained on fully mixed songs, so that's all it can generate. High quality stems would require it to be trained on stems in the first place, which would probably require more compute power. I'd like stems, as would a lot of other people, but I don't think we'll see true stems for a while.

[elfin echo]

You are right about using more compute power, but I disagree about having to use stems for training. Its advance enough to add perfect instruments to trap music with same flow and groove where I didn't see anyone do before. I made perfect transitions using folk accordion in second prompt.

[reef marten]

can we make our generations private ?

I kind of don't want the zero effort continue on stuff i'm going to use, because it's a zero effort add random lyrics to my songs and I know what people would do on the internet.

[hybrid crane]

Yes, you can make your creations private in Suno. If you use the web app, they're private by default unless you change them to public.

[reef marten]

ah, good to know. If you're any "based" rooms, you know what i mean

[robust sparrow]

I love Suno. It's great AI! But what I would really like if we could select the generated image before publishing. The final image often doesn't capture the vibe we want to express with the image. It would be good if we could choose one of the generated images that was created during the song.

[rocky night]

@robust sparrow what would be better is if you could upload your own picture or just generate that which wouldn't be a problem

[void mantle]

Here's a couple of my feature wish-list that I think would greatly improve the Suno site design:

• Show the username of who generated the song. (with possible link to profile with more songs by user)
• Loading more pages for top songs and new songs lists.
• Search songs by name or tags.
• Show number of plays and likes in our generated songs list.
• Show in song data if the lyrics were AI-generated or written by a human.
• Improved cover art images (less birds and rainbow/abstract image bias)

[opaque cipher]

Hi Team, just wanna know if suno can link accompaniments with musical scores and make the scores dynamic or will there be any similar features in the following updates~?

[mighty echo]

Hi! It is there a posibility to make the music with the rythm I want? Like I whistle and turn that into the lyrics I imagine ??

[shell mirage]

Would love to see continuations expanded at least 30 seconds to 1:30.

[brave talon]

It would be amazing to generate guitar chords for the songs to be able to play them

[young ice]

The ability to edit the lyrics on the Suno Website after we combine all the clips into one song would be nice. I didn't know putting all lyrics into one generation then combining with one of different lyrics put them all into one

[blissful current]

I believe this to be an architectural constraint, i.e. can't be increased without updating the model's architecture itself. But every new version of Chirp has increased the max lengths, so I suspect that will also be true of the next version.

[cyan siren]

hello, my account is abnormal

@formal bear Can help me out?

[narrow talon]

Hello, I would like to see a revision of the New and Trend page. Currently the songs only stay in New for a few moments, if they appear at all.
I think it would be better if you could see all the songs from a certain time period, e.g. the last 24 hours. Plus a filter for e.g. music style, instruments, language, BPM or similar to find what you want to listen to. There are so many good songs that were created with Suno, but disappear after seconds and are never listened to. In my opinion, this is a huge waste of potential.

[quasi marlin]

Search songs by tags 😍

[modest latch]

hi how to download the lossless format like wav or others?

[tepid trench]

This isn't available. Maybe in future it will be but the quality of the audio is lower than the mp3 files so WAV would not be useful at the moment.

[wooden fable]

adding further degradation by doing mp3 compression makes it even worse though

[sour loom]

Hi

[tepid trench]

It'd be interesting to see a 1:1 comparison of wav and mp3 from Suno to prove what quality loss there is/isn't. I would've assumed MP3 would be bad but devs have said before that there's no benefit to WAV at the moment and it'd just be a placebo if they offered it.

[wooden fable]

there are lots of comparisons on mp3 vs. wav online. the effect it has on suno's music will be the same as on any other kind of music.
as a musician creating/gathering material you always went the best possible quality.

any unwanted mp3 compression should always be avoided in the intermediate steps of a production workflow

admitted, with the move from v3alpha to v3 it makes less of a difference, but with v3alpha the quality was good enough that i think it was a shame not to offer .wav export.

also, the devs may have different opinions on quality and diminishing returns than someone listening in a sound studio, being able to pick up far more detail.

but i guess that's all a matter of ambition level

is Suno supposed to be a fun toy or a serious tool for professionals. i don't know.

[bold cedar]

Yes please

[tight venture]

My text always isn't followed exactly as I have written it. Are there settings for that? I just deleted 3 "continue from x"-songs because the lyrics were blended together with the previous part although there was complete new text available.

[tepid trench]

Well, the home page has examples such as "make a song about mom's cooking", and there was the Valentine's day site as well, so I'd say it was aiming to be a fun toy. It lacks any tools to edit what's generated, no stem splitting etc. I treat it like something that can give me rough sketches of music ideas, or some samples to manipulate and use, so I guess I'm just less bothered by it being mp3 for that reason. It'd be nice to have stems, wav output etc. in future though

[wooden fable]

fair 🙂

[worn atlas]

Splitting the song into stems as a builtin feature has appeared to be Highly desirable .
Competition is coming.
For now suno's output quality is the best but please dont loose your advanced position.
https://youtu.be/ARGhUD5xSiM?si=wqLGf3sm33UN3GbX

[reef marten]

someone just mentioned crypto payments, and that's an excellent point. we can haz crypto payments for suno AI ?

[hollow locust]

we definitely need 'trees' for 'contine from this song' generations.

[pallid notch]

not sure if someone has already mentioned this but it would be cool to be able to upload an existing song and convert it to something else while keeping the same melody. For example, converting Linkin' Park In The End to classical piano

[tepid trench]

If they were to make this possible, they probably would block copyrighted songs.

[pallid notch]

imo it should be fine as long as you are not making profit off the converting songs

[vagrant escarp]

Splitting the music into stems would truly be a GG moment

[midnight cipher]

Setting Duration for the generations would make it helpful and easy to divide into parts.

[shell mirage]

The ability to hide our prompt words when sharing our song via link or the explore page, would be great for maintaining some form of privacy over how we got our beat/song.

[prime widget]

#💡┃suggestions-web I’m wondering about price structure for the 10K a month credit membership. For (currently) $30 a month, 10k credits are included, but to purchase just 4K more it’s another $30 - I don’t want to open another account as it will mess up my workflow and playlists etc. - I’m hoping you guys can adjust pricing accordingly to fit within each membership.

[hybrid crane]

In Suno's credit system, the 10K credits for $30 a month offer substantial value, but additional credit costs can seem disproportionate. This pricing reflects a balance between predictability for regular users and the need to manage extra resource allocation for those needing more credits.

[potent crane]

no

[steep canopy]

Some feedback to the devs (LOVE this app): I put in "70's prog rock in the style of Yes/Genesis" and the result was more of the "modern" prog style, i.e. a harder more "metal" like sound, when I was hoping for a more classical/madrigal sound. I suppose no big deal but something for future development for a style that the engine can learn.

[blissful wolf]

There are some great AI sites where AI can take an upload of a headshot (or torso or even a full body) pic + text and the site will animate the lips to lip-sync the words in the text.
This could be to added to SUNO's output in a video editor if it weren't such a beach to adjust the speed of the lip-sync to the music lyrics.
BUT
It would be relatively easy for SUNO to match the speed of the lip-sync to the speed of the music lyrics ? instead ?

[sturdy niche]

give us normal commands to tell ai what to keep

[chilly ledge]

Hello everyone! Can you add the possibility of earning credits through views or likes?

[rocky night]

With the countless bugs if you give back say we are refunded or end up with only 5 coins. It should turn into 10 coins the next day. Its insane the bugs we have and it doesn't convert over the next day to 60 or something or make it where we get 100 and then let us test. For a day or 2

[rocky night]

They need more then that. If i was to put in the time i wanted to start from it would glich out and I had one song write something else and even had lyrics in. Its screwy if you keep generating lyrics. Its not perfect.

.
But with this Ai tech... when is it illegal to use this site and one of my songs turn into MGK ripp off and i wrote the lyrics and everything and even sang the song. Soo i dont know. Ai is getting insane. Took me 20times to make my song perfect.

[scarlet sequoia]

Train to do fart noises thanks

[hybrid crane]

(reposted as image to give the team their well-deserved weekend and avoid pinging)

[dry robin]

agree with all of this statement, but i create metal/rock music. ive not been able to create half as good songs as version 2 and 3a

[sturdy niche]

I have more tries to go because it loses chorus because unfortunately i need to continue chorus's end because it was cut idk even why

Instead of minute continuation was 58 seconda

[leaden marlin]

lucky you. my last generation with chorus and instrumental outro got me a 10 second clip with the last 3 words 1 more time 😂

[bitter delta]

Option to choose between male or female singer, also the key of the song should be an option as well

[raw jacinth]

Get whole song to download full mp3 of all the parts mixed together even if they have overlapping parts

[tidal wharf]

It’s exactly what it does?

[raw jacinth]

it is not working then, at least for overlapping ones

[raw jacinth]

nvm, it works, it is just that if you create it from the playlists view there is no feedback that it will create the song on the clips section

[steady scarab]

While Suno quality remains unmatched, the competition is advancing rapidly.

[proud slate]

Hello!
I have a few suggestions that I have and think will be really helpful:

• public playlists
• the ability to add our own pictures to the songs / choose from any of the components that made up the song
• when the video generates / the lyrics are added to the final song, it would be nice if it will reflect the cuts done in the song as now it still keeps the ones from the full uncut song
  On the same topic (but probably extremely difficult to do) an idea would be to add a rescan button that would make an AI scan the song and remake the lyrics of the final song
  Thank you!

[twilit orchid]

https://app.suno.ai/song/5d466dea-6e8e-4148-861a-17d87483faa0/

[hollow locust]

Sure would be nice to have some sort of 'time of creation', or 'created on [time and date]' tags auto-generated on each generation.

[ionic skiff]

please make it easy to implement fade out endings (like we had in 80s songs!)

[zealous yoke]

A few things I would really like to see:

• a way to "upscale" songs to CD quality and export as WAV
• as many others have said, the ability to export tracks as stems
• the ability to keep the music and vocal melody, but regenerate the song with different lyrics
  Also, I don't know if it's supposed to work like this or not, but sometimes I'll only get a clip of less than a minute and it will abruptly cut off mid-chorus, and when I "continue" it, it may again end mid-verse after only 30 seconds or so. I'd like to see some way to indicate that you want it to continue & finish a song.

[reef marten]

1. editing tools. Really needed. There are sometimes radical goofs i want to delete while keeping other things that it generated. Snap to measure on that front would also be nice

2. style change in continue is not something suno seems to recognize. that would be nice to have, but rn it's a box that I fill in that seems to do nothing, on the continue

3. when files are posted maybe the username should be attached to them, so that we can tell our own stuff from other people's

[granite lichen]

Add [Doubletime] plz

[hollow locust]

Something really needs to be done about the lyrics in 'continue from this song'... It visually becomes quite messy!

[reef marten]

uh oh, my 1890s instrumental classical is all turning into modern music with basically no content

I got one generation out of ten, the other nine were 2 minutes of staticky nothing

oh, my bad, im on v2

[hearty peak]

Hi can you make it so suno allows us to choose the key, beats and the chords like:

[Am] [4/4]
As I walk down the road
[G] [4/4]
I saw a cat

or

[Am] [G]
As I walk down the road

(when road is sung, it would change the key to G)

thanks!

[hybrid crane]

Hey Dana! That’s a cool idea, but in 9 out of the 10 times specifying keys like that doesn't work. It’s more about guiding the overall mood and style rather than dictating specific keys or chord progressions. Keep experimenting with different styles and moods for unique creations!

[glossy cape]

So after using Suno for multiple hours straight with no intent of ceasing, I am enjoying it plenty. Two things on my mind that I think would make this a bit better is longer character counts for tags since I can only input about 7-10 before reaching the limit. And maybe some mechanic or system where you can have the AI prioritize instruments, vocals, or other auditory themes. Because every now and then it'll minimize an instrument or sometimes not include it at all.

The latter issue isn't really an issue but it happens every now and then.

[lament crow]

It would be great if users could upload a snippet of music and the AI could analyze it and provide keywords associated with that style. That way, users don't have to waste a lot of credits trying to figure it out via trial and error.

[glossy cape]

Someone actually just told me a site that does that for you, you'll still have to type in the tags manually and it isn't related to Suna, but it's worked for me.

[unborn remnant]

In the custom create form, having an auto-suggest (autocomplete with some suggestions) in the genre/style textbox would be a nice enhancement. Also when using the random button for styles it would be nice to be able to navigate the history of that text box. Sometimes I just click that random button a bunch of times and end up thinking some previous suggestion would be interesting to try but it already got overwritten.

[lament crow]

Found this one: https://sonoteller.ai/pAgnJDJN4VA

[glossy cape]

Yeah, that's the one.

[lament crow]

Please add a 'Cancel Generation' somewhere, as I've hit 'Create' by accident more times than I care to admit. I'm pretty sure I'm not alone.

[dusty raft]

an improvement like a upscaling option would be really cool

[storm lance]

hey all. i will pay ungodly amounts of money to be able to write lyrics for my beats and have an AI vocalist sing on top of the beat. food for thought

[karmic snow]

Im sure this has already been suggested before but I'm just gonna say it again to put another nail in the idea.

Ability to grab stems from a generated song. Sonauto (another ai song generator that's a competitor to Suno has this ability and I'm honestly shocked this isn't already a feature.)

[rare spruce]

I joined this discord specifically to request stems, and would you look at the above. 😄

[spiral summit]

“Continue from this song” doesn’t always work well, it would be good if the AI could make a whole song from the first scratch we did on v2/v3 - with the same melody and rhythm, being able to change the lyrics for the final version

[random sparrow]

Option to generate only 1 song, at a time.

[hollow locust]

Good point... or more. I mean, 10,000 tokens (for example), and each generation is 5, meaning 10 is spent... Why not just have 2,000 tokens and each song costs 1 generation? 🤷

[faint blaze]

How about splitting the "Lyrics" form field into 2 separate fields in "Custom Mode".
The initial input gets lost after pressing the "Make Random Lyrics" button, but i often want to run my input multiple times to see if i can get something better out of it.
If i forget to copy it at this point then its just lost.
A history of initial input and the created Lyrics would also be nice tho.

[brittle kindle]

An easier way to create & rename playlists?! when you create a playlist why call it untitled by default?
Why do I need to be in playlists to create a playlist? Why can't I create a playlist from the add to playlist popup?

[zealous musk]

A way to tell the IA the song is the beginning or the end cause using [intro] or [ending] don't always work well or leads to hallucinations in the system

[brittle kindle]

I also want clips/ songs in the library to show if its already in a playlist

[brittle kindle]

Bulk renaming, checkbox to select multiple songs (add to playlist, etc), rearrange/ sorting the order.

Having extended songs linked to the original. Sorta like it's nested

Original song 🟩
- extended song 🟩
- extended song 🟩
Original song 2 🟥
- extended song 🟥
- extended song 🟥

Some free credits for my suggestions lol

[hollow locust]

I find [outro][end] (as opposed to [ending]) gets pretty solid results. But I otherwise totally agree.

[runic grove]

DAW pluggin
-generating samples
-custom vocals
-custom sounds for MIDI input
-sound library
-sample library
-MIDI input

and maybe even auto mixing

VST pluggin to be exact

[tranquil whale]

Good morning. I have more suggestions.
1º Being able to save the voice style that you did not like, or being able to select it from the song that we already have made, and being able to rename it for other projects.
2nd Being able to download the tracks of the song separately in wav quality
3º For when the editing mode, in which we can add, delete or edit the elements of the song.
4º Extract midi to work in our DAW

[atomic jasper]

It’d be cool if paying users could still get access to V3 Alpha….even though it is a “step down” and has nothing to do with using more compute/higher quality, your secrets are safe with us.

I and many other paying customers just want to use the crappy version.

[bright flame]

Maybe this is already doable (brand new user) It's be so useful to be able to generate an isolated instrument track. like "an isolated drum track from a vintage motown song" or be able to specify the instruments you want with check boxes or something, as opposed to just stemming out a song that you've generated, which I see a lot of people have asked for.

[dusky charm]

they already said that' s not going to happen. It's sad ... but that's the way it is

for certain genre (basically ANYTHING involving a singular singer singing over an extended period of time) V3 is a disaster

[atomic jasper]

Yeah my singing voices have been sounding all robotic and low quality, even more so after they did an update a couple days ago. So now I wonder if they Nerf’d V3 to save their compute budget because of all the new users.

What bothers me most is I don’t know if I’m just being paranoid and seeing things that aren’t there or if this is real and I’m being gaslit into thinking there is no change. Pretty certain it is real and intentional. Would be cool if they told us.

[dusky charm]

if you check in bugs I posted a jazz song a bit earlier which takes 'advantage' of the fact that you are not paranoid and seeing things. it's super predictable

sadly, nothing we can do about it. I'm still (again) subb'd (premium even) but if it stays this way I'll just unsubscribe and hope the next model is able to produce something palatable again.

[atomic jasper]

Thanks I’ll take a look, that’ll help me know I’m not being gaslit

Reminds me of when ChatGPT made those changes last years to save on costs and made the model worse

But yeah you’d think they’d roll out the carpet for paying members but I guess that’s their prerogative

[dusky charm]

they have currently more customers than they have GPUs available, thanks also to an article in Rolling Stones, which, ironically, mentions a generated mississipi Blues/grasroot song ... something V3 is absolutely incapable of (but at which V3a was quite good). As long as people stay with hiphop or pop the vocal deficiency doesn't really matter much. Most pop is already so heavily autotuned and uses so much reverb that it DOES indeed sound good.

"good" not being about the song itself, just about the quality compared to what plays on the radio

[hollow locust]

ChatGPT doesn't quite understand how Suno.ai works, and often generates things like:

[Upbeat, energizing instrumental intro]
[Instrumental bridge with a soaring melody]
[Final chorus with key change]
[Strings swell into an energizing beat]

and I honestly think implementing those kinds of instructions would be awesome!

[errant orchid]

Hi can we have a search function in the Explore tab ? to search music by genre/ theme?

[plain cosmos]

See here they should be ble to scale up. https://modal.com/blog/suno-case-study

But looks like they are over the peak.

[hard tulip]

SSStrongly agree！！！！

[dusky charm]

yeah ... afaik they do, which is why the 'currently only available to paying customers' banner isn't a permanent fixture

[fathom sluice]

Ok, you have built an exceptional product in SUNO, here are some suggestions to make it awesome for artists, especially those, like me, that can sing and play an instrument, but are not good at creating great melodies, working out how the phrasing of words will go to the melody, or creating interesting song structures. SUNO does a brilliant job of those things. However, on behalf of aspiring artists, here is how it could better help us:

• being able to specify/change the key of the song.
• being able to specify a range of notes for the melody to be in. (this would be awesome as then I could create amazing songs/melodies in my range and sing them)
• being able to specify/change the tempo of the song.
• being able to change the lyrics of the song after it is created and have SUNO recrate the exact same song, but with the specified lyric changes (could just be a word here and there)
• being able to export the song with each instrument/vocal on a separate track. This would be amazing as I would be able to import the tracks into my DAW and then toggle the vocal on to practice the melody and then turn it off to record a track my voice against the rest of the song.

[swift jetty]

add a feature like a on off button to add intro and outro to each song intro on or off outro on or off button

Also add a button that is very visible that allows us to see all the commands we can use to shape the song for example
use (yeah oh oh oh) to hear that in background or hear it as effect stuff like that

[twilit ice]

api (maybe paid for customized version or something like that)

[rotund orbit]

How about a prepend based continue option?

I would use something like that in all kinds of ways. Like making whole songs backwards, or adding intros

[tight willow]

Being able to re-use a seed is a must. Kind of surprised it's not in there.

[tranquil whale]

I use MOISES.AI to extract the stems of the song, but it loses a lot of quality. Until they enable downloading by tracks in SUNO, there is little we can do

[dark notch]

nearly all songs have no 16k - 20k range, in v4 i would love to see a complete range of audio. it´s like buying a car with 3 wheels

[cunning tusk]

following on from this about the only thing I could add would be having the vocal options in a separate entry box and having the ability to re-do the last track but only change one option.. incredible app, thankyou!! 🙂

[compact lark]

👎 suno is getting worse and worse with each version.👽

[vocal fossil]

#📜┃rules

[warped finch]

uploading existing songs and extend them (like jukebox ai)

also you should be able to download vocals and instrumentals separated

[pliant shadow]

im happy with the songs it makes

https://app.suno.ai/song/b7b1e3f9-f488-4a74-91b9-f7f206f598a8

[pastel lynx]

How can I fix the ai wanting to repeat a beat constantly and come up with nothing new? I continued from 16 seconds but it's like v2 and the worse Daft Punk songs. I do like the main riff though. edit: sorry wrong thread

[urban hollow]

Please urgently implement the option to subscribe via PayPal.

[pliant shadow]

maybe try to tell it to not repeat a beat constantly

i dont know much about the ai but i think that would probably help

yeh

[spice glen]

a song competition where mods judge them and the winner gets a few credits

[stuck field]

Please add a SEARCH field to the user library. After generating hundreds of songs it is difficult to find old songs, even if you know the song title.

[desert whale]

Anyway to bring back v3 alpha as an option for paying subscribers? Finding the output to be slightly less powerful in normal v3 🤔🧐

[desert whale]

Plus one for individual files, (if that’s possible)

[warped finch]

let'say winner gets the subscription

[scenic burrow]

I keep finding myself visiting suno trending page daily. But you put only current trending music there. Maybe you should add ability to listen to trending tracks from past days. Lets say music that passes certain popularity metric.

[leaden marlin]

My definitive wishlist for Suno AI:
• add or improve genres such as pluggnb, new jazz, jersey club, hyperpop and rage
• being able to import voice models from RVC and being able to create songs with those voices, also being able to make collaborations between the voice models when creating a song, such as putting several voice models in a song and choosing which part each one sings, making some parts in choir with various voice models
• choose in each part of a song whether to sing or rap
• be able to choose the exact length of the songs up to 5 minutes or more
• being able to choose how you want the intensity of each part of a song to be (example: the intro with low intensity, the chorus with high intensity, the verse with medium intensity) and so on with each part of a song
• being able to choose whether or not we want the song to start with the melody (as usually happens in real songs) and not suddenly start singing
• that songs can be downloaded separately by vocal and instrumental
• be able to make songs with explicit lyrics
• be able to upload or link a song to copy the style
• be able to train suno with your musical preferences as if it were a LoRA
• best quality music and voice

[grave pecan]

• Why does continue only ever do ~1min instead of continuing to 2min like the first part? I am wasting gens having to continue multiple times.

• How do I limit gens to 1? I don't want two gens every time

[vale rose]

my current main suggestion would be to have a function that allows you to create songs with the same voice model. I think render ai did a similar thing with cloning/keeping the same face of the person, while simply changing the expression, posture, and background. I think that would be really cool, so that way you could write entire albums with the same voice

[calm oar]

One suggestion, have the option to create really short jingles or audio logos, would be awesome.

[visual brook]

Is there a feature request yet for a place to see every song I've ever Liked yet (while browsing New and Explore)?

I would like to see my Likes

[loud needle]

More than 1 min when you continue a song

[plain cosmos]

vote for your issues here: #🗳┃mini-polls message

[winter wyvern]

ability to provide (original) instrumental audio files and have chirp sing on top of it

[sturdy vapor]

How many people gotta ask for V3Alpha back before Suno gets it? It was so bad I had to dispute my annual membership

[ionic skiff]

after my continuous experimentation and about 500+ credits spent, i think we need a dedicated functionality in Suno.ai for implementing fade outs - the prompts simply do not work

it would be nice to have a tool that allows us to insert fade outs a a desired point of the sound clip, within suno.ai

(its very inconvenient to do this in ANOTHER tool)

[sly kindle]

One feature I would like is equivalent to image in painting. E G we can upload our own backing tracks and have suno play a solo or write vocal tune over it

[azure mulch]

Could we specify the language of the lyrics? Sometimes the "same" words have different pronunciations in different languages, such as Latin languages. Another example, a same Chinese character having completely different pronunciations in Chinese, Cantonese, and Japanese.

[lavish geyser]

A way to trim or crop a song would be nice. If it exists it is hard to find.

[vale frost]

A few suggestions:

1. I feel like the lyric generation, for when you just want Suno to generate the lyrics, creates songs that are a little too long for the roughly 1 minute 30 seconds songs it creates. Having less lyrics leads to less compression and "AI" sounding voice trying to get squeezed into a short amount of time.

2. This is definitely more easily said than done, but I feel like there is a disconnect between the time signature of the instrumental generation, and the vocals/lyrics. If there were some way for Suno to better tell when to extend or shorten the vocalization of lyrics, I feel like the audio quality would increase.

3. Please Please Please give us a search feature!!! I want to be able to see more than just the 50 or so songs on the new or trending tabs. Also, going off of what some other people have already suggested, public playlists would be great too!

[desert crest]

Been playing with the program for a bit now. One issue iv run into is find the PERFECT backing track, but the lyrics are messed up, or maybe id like to add my own vocals over it - It would be awesome to be able to get the Instrumental Only and then the Instrumental/Vocals files.

Iv tried to continue the song with no lyrics, tried the remix function. but cant seem to find a work around for this

Thank you! < 3

[signal tide]

I would love there to be a more expanded explore page. Once a song leaves new, it's effectively hidden even if it was gaining traction. Would love a feature to search for public songs and sort by genre tags

[chilly ledge]

Up.
Also put the possibility to finish non-commercial songs by using premium credits to earn the rights

[leaden marlin]

I am having trouble getting the bot to make songs about some bible verses. it flags them. ( violence). The bible of all things shouldn't trigger it

[bright socket]

I need a way to pay for it without an credit card....... in my country credit cards are rare we use debit cards,paypal and so on.

[unreal minnow]

please make a feedback showing what badword you are using... im receiving a lot of warning saying "Prompt flagged" but i dont know what word im using wrong '-' im not using any "badword"

[plain cosmos]

and did they refund? I tried the same. Not enough people vote here: #🗳┃mini-polls message

[unreal minnow]

yeah, refunded

[merry thunder]

it would be really really nice for further producing, if the model would work like, generating each channel by its own and then mix it finally together and give you the option to download one or all channels seperate. THAT feature would be the ultimate killer!

[glad citrus]

my suggestion is to make it so that when we make a part 2 and/or part 3, etc to a song, when we play part 1, it will seamlessly autoplay to part 2, then part 3, if autoplay is enabled

[lament summit]

I would like to have more control over the song, for example instruments, vocals, solos, intensity, etc.
Can this be done in the future?

[signal tide]

Also random second suggestion: use OpenAI's embedding api to set up a vector search that allows users to search for songs based on the meaning / vibes behind the lyrics and song style

[worthy cave]

It would be amazing to have a radio feature, find similar songs from other users based on the one you create, or any song really. need ways to explore creations!!

[upper inlet]

Maybe have a feature where we can report what we didn't like about a song, like it ending too soon or switch genres mid-way through the song.

Also song outros for the music

[sonic laurel]

exclusionary instructions

[calm vapor]

a list of public songs + search feature from that list would be nice

[lament summit]

Sometimes the song starts very abruptly..

[lament summit]

Maybe with a [introduction] statement it can be corrected.

[zinc vessel]

How can I create a continuation of a song saved into my hard drive? Or is that not implemented yet?

[grave pecan]

Please make it possible to change your profile name on suno.ai. I signed up using my gmail and I don't want my actual name visible on my public shares 😛

[hybrid crane]

What makes you think we can see your email or name with public shares?

[grave pecan]

it uses my gmail name as my profile name on suno

that's the problem, I can't change that

It's importing my real name via oauth

[hybrid crane]

But you say "I don't want my actual name visible on my public shares"
That is not showing when you share.

[grave pecan]

I'm not going to show you it, but my profile name on suno.ai is my actual IRL name of my connected gmail account, and there's no way to edit it on suno's side. Usually when site's let you connect via third party auth, they still let you pick a user name

[hybrid crane]

But nobody see your username, nor email address.

[grave pecan]

Still, don't feel comfortable that will remain forever, I don't like that when I look at my profile it looks like my account is my IRL name. What if devs decide to show profile name in explore? Than I am SOL?

[hybrid crane]

That would be against the privacy laws of Europe.. so I doubt that.. but we will see 😄

[grave pecan]

Yes, and I am in europe, hence why I am even bitching about it 😛

[hybrid crane]

Hey.. me too 😉

But its a valid point.. I will give them a heads up about it.

[lament summit]

Hello Daizy, I would like to have more control over the song, for example instruments, vocals (female/male), solos, intensity, etc.
Can this be done in the future?

[unborn sedge]

Will we be able to edit the existing track ? The Ai mispronounced some of the words. I just need to change a few words, the instrumental is nice so i want to keep it as is

[hybrid crane]

I am just a moderator, but yes.. the Devs do intent to make their product better with every version. How much better is like asking when it will be raining again. 😁

[lament summit]

Thanks for answer :3

[heady sluice]

i will re send this message because its pretty important
Splitting the song into stems as a builtin feature has appeared to be Highly desirable .
Competition is coming.
For now suno's output quality is the best but please dont loose your advanced position.

• A mastering option will be awesome
• voice ai tuner
  https://youtu.be/ARGhUD5xSiM?si=wqLGf3sm33UN3GbX

[grave pecan]

FYI you can do splitting locally using UVR: https://ultimatevocalremover.com/ I use it all the time for mashups and extracting stems

[hollow pilot]

is it possible to download a .wav file instead of an mp3 of the song your create?

[lofty sonnet]

Unless I don't know how..... Could we also be able to rearrange order of Playlists? Helpful when adding to playlists to have more organizational options

[lament crow]

Please MOVE the red notification box (when deleting an item) out of the way so I can access the play controls!!

[ashen robin]

Any plans for an "in painting" like feature that the image gens have? I am thinking you can highlight/select from timestamp 1 to timestamp 2 and then do one of the following options:

1. Remove: This will remove that section of the song, but make sure there is a clean transition from the 2 timestamps (this might require adding a few seconds of instrumental to ensure the transition sounds right)
2. Replace: Add some lyrics (and style like normal). Make a small generation that starts where timestamp 1 is and cleanly merges into timestamp 2.

This would be a very powerful feature if it was doable.
Also, the timestamps would be great if we could get down to the milliseconds (not just straight seconds).
There are times when I just want a word or 2 replaced. Pretty much inline editing. I know some of the voice-only apps have features like this.

Anyway, keep up the most excellent work! Suno.AI is awesome, powerful and fun to use. Thanks!

[wheat crow]

Could you install the navigation system? Because it's very bad not to have this system, if they put this system in place it would help a lot!

[grave pecan]

Can we get negative prompting? I really want to put vocaloid in the negative prompt 😛

[karmic snow]

Ability to save a songs "seed" to reuse the songs rhythm and vocal patterns for a different song. Basically allowing you to redo/remaster a song while still keeping the song the same without changing anything except the lyrics

(Its incredibly dissapointing to find an amazing rhythm just to never see it again if you messed up on any part of the lyrics or just simply forgot to do something with the lyrics)

you could have individual seeds too, one for the vocals and 1 for the instrumentals

[sudden arrow]

Hello, good afternoon, I am new to suno.IA but something that could improve the platform and the way in which you can earn more money is that you can enable a cash payment option to buy subscriptions, the truth is that there are not many people who have one. credit card on hand at the moment, so I propose that a cash payment method would help a lot when purchasing a subscription.

please 🥹

[iron harness]

The ability to continue using a clip is great but could be improved. I would like to be able to combine clips without the melody changing. Ill make a song I like and want to build on it but anything I add doesn't match at all. Maybe its me....?

[leaden marlin]

Allow one generation instead of two so that credits don't get used up twice as fast

[chilly fable]

There is a channel that distributes sound sources generated using suno.ai without permission.
Please restrict this channel. This is a clear disregard for the rights of other subscribers. youtube.com/@Free-BGM-Library-xe2lu <-this channel

[clear chasm]

as long as you're a premium suno member i think its ok

[sweet rune]

Editing page

• editing orginal song
• add or remove intrument
• add eq or filter and effect on some segments
• add or remove lyrics
• editing part 1 2 in different order
• rythem generator from recording so ai make music base on the rythem you make

Option to allow cobaration
Same tool with above and can't dispute without permission

Our own music streaming app

• seamless upload to steaming acount
• collaborator get credited when upload to streaming app with tag link to they acount
  -monetize streaming by allowing ads to be on your track and get pay in crypto or cash

Steamer can also earn crypto by

• creating a playlist, the play list that get to top leaderboard will get a share of the stream revenue

[sweet rune]

Yes we need this ASAP

[fiery herald]

Song contests!!

prize being credits or cash 😄

[zinc vessel]

I'm going to ask this again, since I'm getting ignored.
How can I create a continuation of a song saved into my hard drive? Or is that not implemented yet?

[crimson trench]

Hi, musician here. I'd love it if suno could try and complete tracks I started myself

Just the idea generation and inspiration alone would be worth the entry price

[zinc vessel]

Just like Alan said. I think it would be very helpful! 🙂

[lilac lake]

On Instagram people are still saying they'll make songs and share them on various platforms without getting a membership and earn money from them, why aren't you doing anything about this? And why is v3 public? Then why am I paying?

[worldly galleon]

When you click 'delete' on a song, the big red popup window that lets you undo it shoudl not appear over the song controls at the bottom

[zinc vessel]

You're paying for early access for other new modules and for more credits. 🙂

[lilac lake]

Then I don't need to make another payment because there's no advantage, innovations are already coming to the free membership in a very short time anyway lol

[zinc vessel]

Oh, really?

🙂

[sweet rune]

What do you guys think?

[reef marten]

the suno copyright on free vs paid accounts is super ambiguous, and this ambiguity will be a nightmare for suno support, and long term, maybe it's legal team.

Much better if each song has an icon showing the ownership of the the song. Also, it would be nice if there was an automated way for users to gain copyright control of unowned or suno owned material that said user has generated, or best if they automatically gain that ownership by sending an account to pro or premium.

Additionally, these rights should track through "continue" (but not "remix", since an AI prompt cannot be copyrighted) so that some other user cannot just continue someone's song and add a null entry file at the end, get new whole song and then claim ownership of the song they added nothing to.

just all this sort of stuff should be automated and clearly marked, or it'll be a total headache for suno team to deal with when there are a million users

[wise coral]

I'd love the ability to upload vocals in my voice and then the AI adding prompted instruments to it

[lilac lake]

Yes, the problem is that currently I don't think free users and paid users will be able to distinguish between them. If someone says they can understand this difference, then those people can even see Smurfs. If they really want to control it, they can do it like Envato, where if YouTube or another platform notices it, they ask users to enter a code. This way, it can be prevented, but it seems more difficult. My suggestion is if there is a new update, only paid users should use it, then it will be fairer. If you want to give the innovation to free users as well, just grant rights for two songs, not unlimited. Otherwise, people will take advantage of this.

[chilly fable]

I agree on how to create creations by using music for your content or processing sound sources through remixes or vocal recordings.
However, there is certainly a problem with the way YouTube is operated by directly uploading the generated sound source, generating profits, selling it for money, or receiving requests and producing it by proxy.
This is directly stated in the license, and users who try to interpret it freely and ignore copyright or other subscribers' rights will also be a problem, and if we neglect it, that will also be a problem.

[broken fable]

Does anyone know if there are plans in the future for vocals to be applied to an instrumental track?

[nocturne phoenix]

We need a better Discord to talk with the community! Everyone seems to use #💬┃general-chat for everything, and it's hard to follow all the good stuff happening in it. Could we have new channels that clearly define the topics and mods to enforce? For exemple:

• Music feedback requests
• Tips and discussion about lyrics and voices
• Tips and discussion about instrumentals
• Tips and discussion about mastering
• Tips and discussion about music video
• Tips and discussion about distribution and IP rights

Etc... Thank you for this amazing app ❤️

[trail slate]

It's gonna be good if Suno allows to download lossless audio with some amount of credits, for instance, 100 credits for mp3 320 and 200 for wav.

[karmic snow]

ikr? I could actually make amazing songs if I had that ability

[karmic snow]

you can already download it as an mp3

[atomic sedge]

Not sure if this suggestion has been made already since I'm new...

If we can have a way to download stems from our creations, that would be great.

For example:
I create a song with many instruments used, but I only like what the creation did with the drums and bass, and don't want any of the other instruments from an already made creation. With stem downloads, I would select the stems for the drums and the bass (not selecting stems for the guitars or synths), and download them separately.

[leaden marlin]

first thing i thought when using the demo is .... #1) how can we render in .wav files because the 192kbps seems low quality .... #2) can the sounds be separated by track for production usage.

[paper saffron]

Looking for some help here. I have created a short clip which has the desired beat and vocal styles, but the lyrics need to be changed (not the full song + need to improve rhyming)

How do I switch out the lyrics and keep the beat, and make it longer to last the full lyrics? Thanks in advance!

I heard that this is currently not possible, but would be great to see this feature in future updates!

TL:DR Feature request: Suno totally nailed the beat and vocals but I just need to switch out the short lyrics for my own full version and make the beat last longer than two verses 😂

[onyx wigeon]

Why It's generating only 20, 30 second clip when I am using "continue from this song" failing to read lyrics in continuation?

[spare shadow]

#💡┃suggestions-web want to add some animal or some object sound in my song.

[reef marten]

yes, this was my point, if there is a preference to paid users, there needs to be an ownership icon. also, that should stay with the original creator, and the person who continues shouldn't be able to become the owner of the song you created. (which according to the system they totally can right now)

[tidal wharf]

with the playlists feature it would be great we could shuffle

I have a playlist of 1500 generations I need to look through (had too much credits hanging at the end of the month:)

I added them all to one playlist (btw I don't think all of them were added, perhaps there is some hidden limit)

[reef marten]

I see the v3a model fans issue: non-dominant voices on v3 automatically go to choir on chorus and there's not any way we've found to stop them from doing so. v3a should probably be an option on the drop down along with v3 and v2. If there's some system specs issue of the model taking more GPU time then just charge a different amount for it, but I see their point, there are definitely a lot of times when you would want to have a solo voice in a song chorus.

[stone solar]

it would be (incredibly) useful if .srt's were created, even if they aren't 100% accurate for the lyrics, maybe as part of the .mp4 download/creation

[reef marten]

yeah nees a search function also

[trail slate]

mp3 192 sucks with phase issues and lower quality

[trail slate]

In my opinion stems may not work atm bc the machine creates a whole song directly, that's the way they trained it

[karmic snow]

192? I do agree that we should be able to grab the song in a .wav format

[atomic sedge]

right. thats why its a suggestion to add at some point. 👍

[reef marten]

the song is a mess of continues tho. it's not going to duplicate, and i'd need to clean it up a lot. but there are a lot of duets. it's just 2:30 am. i can't right now. but they exist. Suno docs say "this is a bug not a feature" so it's going to depend on a lucky roll. I think you might have luck with [character] prompts. I actually got a dog barking today from saying
[dog] woof
this however, did not work
[cat] meow
it was a woman saying meow, but the dog was a dog.

[man] hello
[woman] come in
might work, but idk for sure ?

[reef marten]

and midi

[karmic snow]

yea ive seen a lot of people say that too. personally idk what id use a midi file for (except maybe for Minecraft) but hey im new to music making so yea...

[karmic snow]

A way to manipulate the speed of sung word/s with a simple symbol prompt before and after the desired word/s. Depending on how many of that ">" or "<" you use (must have the same amount on both sides) the faster or slower those word/s are sung. As it is right now there is no way to force the manipulation of words so having this suggestion as a feature would give a lot of control over how someone wants the lyrics to be sung, right now though we just gotta waste credits and hope it gets at least close to what we want a song to be because I doubt anyone has gotten that "perfect" sung lyric exactly how you imagined. This could progress with more than just control over the speed but also how a word is pronounced. I'm not big brain enough to figure out how that would be done but im sure someone could figure it out. Maybe ">" and "<" are good choices for this prompt feature but I don't know, I think that should be up to the developers.

[nocturne meteor]

add a remix feature for nightcore jamming and fun, must have 2 songs and A.I smashes both together

[leaden marlin]

option to pay with paypal would be great for me! and as previously mentioned .wav formatt the 192Kbs is too low of quality.

[weak dagger]

a way to download a .srt file to help with captioning videos

[dusky charm]

great idea

[violet crest]

💡: When you “Continue this song”, it create an Indentation that includes all parts.

[leaden marlin]

Why is the duration of music creation limited to 2 minutes? Also I try to create a Sonatina Piano and the AI cannot finish but cut the music, extremely fustrating.

[open pasture]

You should create an option to output stem files for movie scores. That will draw the attention of movie producers.

[misty blade]

users channels - each time I’m entering with link I want to see other works of the same user. Of course show only the public ones

Also give option for comments

[vague harbor]

Im sure this has been stated but I am adding my name to the chorus:

-1- i would LOVE if i could mix stems. So add one part from one stem I liked with another piece. A Frankenstein mishmash of my favorite parts. Would be great to see like a board for each song with all the parts. So we can visualize the song we are creating.

-2- a download all button. So i can grab all my clips at once. Preferably filtered : liked clips, full songs, or all.

-3- not sure if my clips are totally private if I don’t press public. I would like that because I heard people are stealing songs and with copyright issues the way they are with music, this is important. Also some form of tagging DRM for the owners name to help with this.

-4- .WAV format or midi(or both!) is NECESSARY. Absolutely no doubt about it. This needs to be worked on asap. I can’t believe us paying customers don’t have this already. Please.

-5- Some form of upscaling or noise cleaner for the songs that have that weird background. I went on lalal.ai and I found the noise to be trapped in the synth part. When it separates the songs. So it shouldn’t be so difficult to clean out with a good engine.

-6- Oh can the Bird just not always play guitar and look so sad? Can we change the song photos from the start to keep the songs all organized by photos?

-7- would be awesome to have an RSS feed for my website of songs I make. But only full complete versions.

-8- i beg of you…GIVE US MORE CREDITS PLEASE. please please. It gets so much wrong. Each song takes soooo many credits to complete. I went through 2,500 in two days.
**
Hands down SUNO is
the 👑KING👑 of AI MUSIC
🎶🎼🎵🎤🎧🎸🥁🎹🎺🎻!

I didn’t think it was possible, but here we are. Wow. Great job. Thank you! I have so many loops and clips now to make hundreds of original songs!** also it is very addicting. I have slept very little this last week. Lol

[vague harbor]

Even if we can just download all the clips I could mish mash it together myself on logic or any music software daw

[dusky charm]

because, like any AI, Suno works with a context window. At some point it loses context (forgets what came before, basically) and starts hallucinating. My guess is that 2mn is a safe value. When you do a continuation it uses the last 1mn of the original clip as context

[dusky charm]

the clips are generated as "one', not instrument-wise

[viscid cloak]

Suggestion:
The option (as a paid user) to disable the remix option, if you dont want your songtext used. I have many songs where i put work into the lyrics. It would make sense, since we cant use texts from other songs or artists

[reef marten]

way easier to edit a midi file in a daw. musically speaking.

[reef marten]

yeah, something like this would be nice. if it was separated out into layers where there was a generated instrumental then maybe markers could go down for where words would go but that's a ways away

[upbeat dew]

im super young and not smart so im sure this is probably a stupid idea or has been already done, or cant be done, but i like to think abt if there was an option to choose whether the voice singing will be a male voice or a female voice i think thatd be kinda cool

[spice glen]

all songs are private by default already

[gusty silo]

could do with an additional button to turn outro on and off. trying to continue a song for another full min and its doing 7 seconds because "it" wants to do an outro.

[nimble kestrel]

Create Playlist from add to Playlist, super annoying that I'll find the tracks I want, after clicking outage by page, then readjusting u didn't make a Playlist first, then on top of that after making a Playlist it kicks you back to the library so you need to again click Playlist, then new Playlist,

[nimble kestrel]

I'm not sure if Possible atmbecsuse of the way the songs are written. Im pretty skilled at stem seperstion and cleanup both Ai based and manual using spectrallab and rxstudio, and something about sunos files really make them bug out

[desert whale]

Is there a way for us to go back and edit songs? I.e say the song is 99% but requires certain lyrics to be re-sung. It would be awesome if we could implement a “rerun” feature that’s similar to the “continue from”, and then we could highlight the word that we wanted to change with a <theword> or something of the like. That would be a game changer but yeah not sure if it’s possible rn due to how the generations work.

[desert whale]

It’s gotta be one of the most requested features :p problem is the audio isn’t really made up of stems as it would be imagined. You can take the mp3 file and extract stems in apps like fadr but you will find that a lot of the audio information of the instrumentation is in the vocal, or conversely if you mute the vocal the song loses a lot of information that would otherwise be in the instrumental will sound like it’s missing.

[gusty silo]

you can specify male/female voice in the style section, it doesnt always work though. eg: https://app.suno.ai/song/074f5341-5056-43a8-bc9d-b3588de57320

[hexed ferry]

+1 request for the ability to export stems. Would be a game-changer for music producers!

[leaden marlin]

Do we always have to click refresh to see new credits or can the website automatically refresh by itself like Twitter, Discord, YouTube, etc?

[ionic dawn]

I want to write a song with a male and female singer singing a duet, but the FAQ says that it doesn't support songs with multiple singers. But it seems like it supports multiple voices of the same gender, like girl groups and boy groups. Does anyone know?

[deft crystal]

have it to wear when you make the song have one with voice and without that would be great c;

[dawn nebula]

#suggestions please make the ia take reference from artists on top of spotfy. I've been trying for 3 days to make music similar to MC Ryan and Mc Ws da Igrejinha and nothing similar to Brazilian funk comes out.

[lament summit]

Can I define the voice (male/female)?

[ionic dawn]

when you make a song, type male vocal under style of music

[grave pecan]

• +1 to request to rerun a gen with the same seed after fixing/adjusting vocals
• +1 to wav export
• please give us the option to add a description or more meta data to public songs (e.g. lyrics author, etc.)

[spring silo]

Request for Feedback from Free Users (Basic Plan) ✨

Thanks everyone for all of your suggestions so far! 🙏 I am a researcher at Suno, and I would love to hear from some Free users in more detail about how you're using Suno so that we can make it better for you! If you're on Suno's Basic Plan, please send me a DM with your Suno email, and we can setup some time to chat ($20 gift card for 30 mins of your time). Thanks!

[celest zealot]

Hi all! My suggestion would help a lot in training to sing a piece or a choir part: It would be great if you could add the possibility to import a MusicXML, and Suno would sing the vocal part. I think it would even consume less power than creating a full new song, because we only need to generate the singing voice, not a whole melody or arrangement.

[vague harbor]

Thanks i wasnt sure about that.

[fresh lodge]

Hi, not sure if it's already been mentioned but in the same way you can use a --no prompt with Midjourney, it would be great if we could direct the AI away from things with series of --no prompts

[ionic dawn]

I want to write a song with a male and female singer singing a duet, but the FAQ says that it doesn't support songs with multiple singers. But it seems like it supports multiple voices of the same gender, like girl groups and boy groups. Does anyone know?

[vapid isle]

• Voice Option choose Male / Femal or own ogg file ? - seperate instrumetal and vocals for further work in 3rd Party Plugin - "continue a song" should use the last 20 seconds of the Song to reproduce the new Mix ( sometimes it skips parts) - more deffiniton in prompt for Intro, Outtro, maybe also Break, or Pause etc.

more information in the " creation" TAB, the Songs you allready added into your Library should have information in wich Library it is allready, sometimes i added 2x 3x the same Song because i forgott i allready added it. need more overview. thanks Suno is best

[red fossil]

https://www.youtube.com/watch?v=GK4kI0R85A0&ab_channel=EIMA
THIS IS ME DANCING HALD NAKED

[heady cove]

please steal this from the lesset ai Sunauto to Suno ASAP

[hot pier]

We cannot a priori export a song with all the separate tracks (vocals, bass, drums...) you have to go to do it for example on LALAI. This is a bit stupid since SUNO, which creates all the tracks, could offer a global export or an export by track. Is this planned in future developments and if so when?

[heady cove]

control track means consistent singing voice across songs

[tepid trench]

Some way to ping moderators might be good, without having to ping them individually. I have no idea which mod to try to get the attention of when I don't see any online.

[old relic]

You are assuming that the AI creates individual tracks and then mixes them, which is usually not the case with music generation AI (I, of course, don't know how suno.ai works, but if it's anything like the usual diffusion or autoregressive models other systems use, there is no concept of tracks internally, and the final wave is generated monolithically)

[warm zinc]

is it possible once a track has been created to ask it to vary it in someway e.g the lyrics, whilst keeping all other elements the same?

[hot pier]

I don't suppose anything actually. it's just a comment that would save us from going to LALAI for example to ungroup the tracks. in this case this functionality which offers other AI could be natively integrated into SUNO which would help home studios in their creation process

[old relic]

yeah, I understand. I think that would be a great feature. What I want to say is that this idea is asked in this room at least once every 20 minutes or so, and most people assume that the AI uses a similar process as a human would follow, and that Suno actually has these "stems", and just not giving us the chance to get them. But in reality, I don't think even Suno has them. So, the way to get this form them I think would be to actually get add a LALAI-equivalent service into the suno website to extract these stems after the music is generated. But you would still not get clean original stems, just the same result as you'd get from an external service.

that being said, I agree that if such service was integrated, it'd be a great feature!

This is the same reason for which it's hard to regenerate a song while keeping one part static. There are no individual parts!

[hot pier]

this would avoid having 2 separate subscriptions and thus strengthen the conversion of SUNO with a complete service (Jacques in marketing leaves this body...LOL)

The separate tracks then allow you to take what interests you and change what doesn't matter. moreover another functionality once the audio tracks would generate the possibility of putting them in midi too

this would avoid another membership to melodyne for example

I think that if AI sweeps the full spectrum of the needs of user creators it's better....

[old relic]

yeah, agreed!

[hot pier]

GREAT IF YOU’RE OK. so when is that due....lol

[old relic]

Another thing that I think would be great for the use case of video game music, but I am not sure if it's possible: can the model generate looping music? (e.g., that the ending matches the beginning, so we can play it in a loop).

[viral ember]

When do we get FOLDERS to organize our Library -_-

[quaint geyser]

Why has both the sound quality and the arrangement deteriorated so much today? It's really terrible. What happened? Please don't ruin SUNO.

[blissful karma]

Ok so I got a great song on the first go round. Like... miracle! But the song is therefore too short. And Suno apparently knows it nailed it the first time because it keeps refunding my credits when I try to continue.😭😂

So I would like to continue by having a spoken word segment and then go back into the chorus and fade out. Any idea how to prompt that?

[wanton thistle]

Mate, my ears are falling off from all the 4-chord progressions deteriorating into granular noise 😭 Thought I am hallucinating, but seems like it is not a one-off experience, after all!

[potent rain]

You need to continue from earlier in the song. If you continue when its fading out or silent it thinks thats the end and tries to complete it

So instead of continuing from 2:00 try continuing from 1:55 or even 1:45

[blissful karma]

Does it matter if it's not the right stopping place? Will it figure it out?

[lofty sonnet]

I would like to see track organization approaches closer to Traktor than those of Spotify, colors, notes, custom tags, star rating systems

[potent rain]

It’ll do its best. With AI you just gotta keep generating until you find a variant you like

[lofty sonnet]

right now we only have track titles

and no way to search in library other than toggling pages of 20 clips

Also I need to be able to rearrange order of playlists, hopefully a UI improvement than you can scroll down through all playlists, no one wants pages.. even spotify just endlessly scrolls once you get to a discog

[lofty sonnet]

maybe like itunes?

[lofty sonnet]

Inpainting is being researched

[spiral marsh]

Would like to see all songs that are publicly available as well as to what account it is connected? Maybe a way to follow a creator as well?

[blissful karma]

Ok! Better, but still failed to prompt a spoken word segment. Any way to do that?

And as a suggestion: it would be great to be able to stich together whichever parts of a song you want. Sometimes Suno guesses wrong.

[lofty sonnet]

and stream monetization as a thank you for mining/creating the most popular suno songs 🙂

[potent rain]

People have been complaining about a downgrade in the vocals the algorithm generates over the last week. Might be related to that

To stitch stuff together you’ll have to use a DAW

[blissful karma]

Fudge. I'm not terribly good with those.

[reef marten]

time stamp on gens

also, clear model and sub model #s every time something changes on the back end. like 3.001

[cobalt copper]

https://app.suno.ai/song/eb6f1e73-2930-419e-ad6e-0402bb83dddb can someone please share the prompt for this?

[reef marten]

you're in suggestions room

[potent rain]

┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴┬┴

[reef marten]

yes

[potent rain]

Thats the prompt

[reef marten]

just click remix

but i think this channel is for suggestions to suno team

[cobalt copper]

oh sorry

[potent rain]

To stitch the suno parts together you can click “get whole song” in the menu for the song, but anything more custom will have ti be in a DAW

[cobalt copper]

yes thank you, but in that one, i am particularly asking for words like which kind of genre or instrumental music is

[feral shore]

In subscription are there other payment systems?

Because i don't have a credit card.

I would like to pay on paypal or something that doesn't require a card.

Can anyone pls advise?

[hybrid crane]

You could think about getting a "prepaid" card (also known as a debit card). There are plently of free ones out there. I am European and have it too. They function as a credit card, but are cheaper (often free as said) and you have total control. You just put money on it, and you are ready to go.
I do know Suno is working on more payment options, but I am not sure when it will be in place.

[feral shore]

Thx for advise ☺️

[gilded igloo]

Hi guys my first post in this discord, nice to meat you all. I have a question: How would it be technically possible to blend two existing songs together, is that something suno is planning? Because I think this would really exallerate creativity

[hybrid crane]

Suno is about creating new music, not blending existing songs together.

[gilded igloo]

well that would actually create new music

ofc I see the issues with copyright

but sooner or later a solution for that will emerge

Imagine mixing a beethoven Sonata with a beatles song

[crisp osprey]

Pretty sure I'm not the first to ask, but it would be great to have the option to create Loops, especially for the instrumentals. (Audo that can be repeated without the end-user noticing it started over)

[gilded igloo]

or any other combination

Ok I might work on that further than ^^thanks

[hybrid crane]

At the moment Suno is text to music. But in the future, anything is possible 😉

[harsh musk]

Is there possible to donwload our songs in bulk? ( all library or playlists ) ?

[viscid cloak]

Suggestion:
Reroll your album cover or atleast when you have 3 parts, choose which you want.
https://app.suno.ai/song/18ecae67-bbf7-4682-a1e8-cb8bf17a4b5b
Its about a shadow Witch and it got rainbows.

[north dune]

Will we eventually be able to lengthen the style of music prompt box? Add more descriptions and whatnot?

[quaint path]

if the song description box could be longer that would be much appeciated. i want to be able to feed in larger prompts. thank you for your consideration. the 200 character limit is quite short

[quaint path]

we were thinking the same thing at the same time. you beat me by one minute

[north dune]

Oh wow small world haha

[atomic sedge]

@nimble kestrel @desert whale
This is the channel for suggestions. I get it isn’t possible now. This is a suggestion for finding a way for it to be possible at some point, like all the suggestions here. Lol.

[frail sapphire]

How to make a neuron sing something like this:
https://www.youtube.com/watch?v=Rs8aq-e_J4M&pp=ygVK0YHRgtC40YXQuCDQviDQvdC10LjQt9Cy0LXRgdGC0L3QvtC8INGB0L7Qu9C00LDRgtC1INC80LDQvdC00LXQu9GM0YjRgtCw0Lw%3D

https://www.youtube.com/watch?v=aPwJ_6HgUTM&pp=ygUf0LjRgNC-0L3QuNGPINC70LDQvNC_0LDQsdC40LrRgg%3D%3D

[flat oxide]

Any plans to implement an audio mixer so we can increase the volume of the songs we make? They seem fairly quiet.

[boreal saffron]

Could the future allow for training personalized models with Suno, ? perhaps through an upgraded subscription ... GPU are not free i know ...

[dapper swallow]

Suggestion to make it so that when you generate a song to specifically choose if you want to generate 1 or 2 samples instead

[hybrid crane]

The devs always said you should see it as 2 for the price of one.

[dapper swallow]

but you can generate less songs then

instead of 10 it's 5

since 5 credits for 1 song and it generates 2 so 10 credits per 1 song

meaning 5 songs

[hybrid crane]

Every part is considered a song.

[sand shuttle]

Can suggestions for the discord be posted here? Or is this just for the AI itself?

[lilac rose]

I want to suggest that the site provide a way to search for published songs by name, style, or mood.

[sand shuttle]

Cause my suggestion for the discord would be a forum channel so people can post the stuff they make and ask for suggestions without being spammed over by music of different types or general arguments.

[summer wyvern]

I have 2 suggestions, the first one is to have a button that shows you the whole list of available styles (or suggested styles based on the provided font), and the other one is to be able to have more control of how the styles 1 influence with ranges or otherwise with priority numbers

Range:
Cinematic: 1-100 Range (30%)
HipHop: 1-100 Range (20%)
Aggressive: 1-100 Range (50%)

Priority:
Cinematic: 1-10 (6)
HipHop: 1-10 (2)
Aggressive: 1-10 (9)

[willow gulch]

Will it ever be possible to transpose mid-song? Going to a higher key, for example for the last final chorus, would be an interesting option

[tidal wharf]

sometimes [modulation] or [change key] works, but rarely

[reef marten]

the seed for the generation is really needed. I had a perfect generation and it got deleted by the spinning circle of doom. I cannot get anything else even close, and i did like 100 tries to get it back

[lapis osprey]

Need to have some way to do "inpainting" with the generated songs. A lot of times I really like the song, but a particular phrase wasn't well sung. I want to select that part and generate new variations that can be inserted in place.

[raven terrace]

I think the most needed upgrade right now IMHO:

• WAV export (especially for subscribers)
  And…even better than stems:
  The possibility to save your “musicians” — for example — use the same vocalist or drummer again, if you liked them in one of the generations etc.

[vagrant crow]

Add crypto payments for users have no VISA/ECMC cards, because MIR payment system is banned.

[agile coral]

Okay, this is actually the kind of thing I feel like I should tell the devs at this point. I lose a shitload of credits and really nice gens because for some reason the AI likes to have lines end on an upward note, making it sound more like a comma.

It's kind of along the same lines of the limitations it has in cramming awkward lines into a sensible flow. It doesn't know that lines shouldn't sound unresolved, it seems.

[granite egret]

For now one can Install audacity and use that for fade outs. Its free. 😛

[sterile pilot]

It would be good to have a search engine for our songs, and to be able to have categories to be able to segment and thus have the songs separated, for example creating favorites, in process. This way we can assign the name to each category and send the musical songs there

[agile coral]

Found a really strong example of a tune that should resolve but leaves you hanging instead.

https://app.suno.ai/song/94a04a96-5c4c-4583-8dcb-29fc4df78645

[ionic skiff]

i know, i would prefer to do it on suno for 2 reasons

1. i am paying for it - fade out is a basic function, should be easy enough for suno
2. suno will suggest dynamic fade out endings, hence offering its creative input with each unique fade out generation - audacity **does not **offer that!

[sterile pilot]

use at the end of letters [outro]

[agile coral]

in theory you could wait until you hear something that sounds like it's begging for a fade-out. Or even not prompt an ending, try to get a vamp that would make sense for a fade-out.

that said

thank you for your good work in debunking a [tag]

that is always appreciated. people and their damn superstition tags

[grizzled frigate]

One thing I can’t get past is the MP3 compression. When I spend 4-5 hours making a song, nail the lyrics, and then parts are layered, the compression is absolutely horrid… like could never be repaired bad in some parts. I would be willing to pay 10x the price to have it output uncompressed stems.

[agile coral]

Does the song sound like that when you listen to it while making it, or does it sound worse when you download it

[grizzled frigate]

Usually it’s slightly over-compressed when making it, but 8/10 time I make a whole song from my parts, there are artifacts that come up where the parts are layered. It’s 100% a problem with the compression algorithm. I actually have a solution that would save bandwidth and do a much better job and could even get Suno in with Dolby on an Atmos plugin I’m developing if someone cares to have a 15min chat.

[agile coral]

damn, a lot of times when people complain about the mp3 compression the answer is "it's not compression, suno just has a noise problem." But I believe you

[grizzled frigate]

The noise is completely manageable in mastering

The compression isn’t - you can’t uncook an egg

I’ve resorted to things that seem completely unnecessary - using tools to separate stems, spending hours converting instruments to MIDI so I can use VSTs, then remodeling vocals in kit.ai… it’s almost to the point of it making more sense to bring people into a studio again to save time over money.

[soft meteor]

I am however unsure if suno ever outputs "uncompressed" audio anywhere in its pipeline. Bark model used EnCodec. I am not sure if chirp does the same. And if there might be a step where it's raw uncompressed audio. But i think it trains on the EnCodec representation

[grizzled frigate]

It could be order of operations… it’s clearly taking shortcuts and layering the vocals and instruments prior to compressing them again. EnCodec should be good for this, but the stems need to be generated separately. I understand the implications of that on resources, but that’s why I say I’d gladly pay a lot more $. They really need a “producer” tier on pricing for this.

[vocal thicket]

it doesnt layer them

[soft meteor]

that would require training on stems.

[uneven hornet]

it'd be nice if "Get Full Song" would let us select which thumbnail to use. the last part of my song is only an outro, so it has a thumbnail that's off theme, but that's the one it uses for the full song.

[ionic skiff]

you mean at the end of lyrics, i should write [outro] instead of [Fade Out End]

[grizzled frigate]

Thanks @hybrid crane - it seems to mostly be in the download that the artifacts really come out a lot more (just the nature of compressing again), but the artifacts occur in generating the song. A higher quality download would be very welcome… maybe charge 100 credits for an uncompressed download or something?

[hybrid crane]

I was just posting that so you could see what the Devs are saying. At the moment it his the best quality they can give. It is how they get it from the model. They don't have an "uncompressed" file. It is a misconception.

In V4 the quality will be better.

[grizzled frigate]

It’s clearly a higher bitrate on the server prior to download, even if it’s not uncompressed at any point

[hybrid crane]

I wish I could tell you more, but all my information comes from Devs and what they have shared. I am a volunteer myself.

[tepid trench]

When you are playing a song on the website, it is streaming the MP3 file.

[grizzled frigate]

So I mean to say without compressing more, I’d gladly pay a lot more

[hybrid crane]

Me too.. I have 2 premium accounts myself 😉

[primal pawn]

After making all the songs, I want you to create a function that only extracts the instrument separately.

[inner coral]

I would be glad to pay 2 x credit in order to have 2 x steps on the generation.

[solid creek]

A function that reads nature / ambience sounds in the "ooh-yeah) fashion .

This is the beach song (ooh-yeah)
enjoying the blue ocean (play water sounds )
and kids playing (play kids laughing )

[spiral marsh]

Option to upload custom art...

[reef marten]

are we down rn ?

[hybrid crane]

The team is notified. They are aware and looking into it.

[tight hamlet]

can you allow us to buy the copyright of a song we made during the free trial? maybe charge like the full $10

[sonic mist]

Exporting video underneath our own art

Like without the lyrics underneath

[desert whale]

Do you know if we will be able to regenerate previous generations with v4 when it comes out? Would be so cool if we could retroactively improve the resolution / quality even just a little bit.

[sour portal]

Being able to Search and Comment on songs would be a neat idea... Maybe devide them into Category / Genre to make it easier.

[nimble kestrel]

Ah my bad I thought yall meant like adding mdx to suno

[jagged fable]

A bit of a meta suggestion for the server: There should be some type of #questions or #suno-technical channel for those who have questions or want to discuss the ways Suno works. Since there is no actual knowledge base or comprehensive guides (besides a small 3rd party wiki), it has been up to users to work together to figure out strategies.

Having all discussions Suno-related discussions solely in #💬┃general-chat makes everything so cluttered & difficult to discuss anything technical. Especially with most people just sharing their songs in there.

[leaden marlin]

Will there be an app for android and iOS?

[crisp ice]

I would suggest adding a folder system so i can organize my songs and files

[half sail]

The following:

• Allow Suno to remember my last settings [ChatGPT/Custom Mode]
• Give users the option to set style presets for themselves

[gusty plinth]

guys. please add vocal seeds so we can have better chance to genereate same vocal timbre

[peak oyster]

There should be an error message (which you can disable if you want) that says "You didn't generate any lyrics are you sure" because I keep forgetting to generate lyrics with songs

[hollow locust]

If it generates with lyrics which start at 00:00 or 00:01, I can't do anything about extending the intro in front of the song. However, if it cuts off at the end of the song, I can just "Continue From This Song..." and everything's fine.

[hoary vessel]

/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.

[north flax]

I think we need fractions of a second to continue at, like 2:01.5, I often can't get the timing exactly right

[feral aspen]

Amen

[north flax]

the dream is to have "in-painting", where you can select a range and replace it, but keep what's after it. I had so many great moments generated but are unusable because of what's before it

[limpid moat]

I would like a refund credits option, lately when attempting the Continue from this song function, i put in 2 verses of lyrics and it will start with the 2nd line of the 2nd verse followed by playing out the song like it's ending. In troubleshooting I even just tried putting in only one verse and it sang the last four words then played out like it was ending. Surely being a AI model if I click a Refund Credits option it can auto check that audio and be like yeah 4 out of 30 words were sang, not what the customer ordered and grant the credit refund. Are credits refunded when you bug report your song?

[solar socket]

please add a feature that always you to use the same voice for multiple songs

[limpid moat]

Thats a good one too like you would just have in the input (Hashtag)Vocals1 or [Vocals1] in Lyrics and it would save the Vocals it Generates under Vocals1 that only is Fetched from your client.

[fiery geyser]

How to delete all records of library in one click

[hollow locust]

It seems that with 'sultry female vocals' it's far more likely to ignore [intro, instrumental]

I'm sure a 'search' feature for 'Library' would be greatly appreciated.

[hybrid crane]

Please do not direct ping staff. Thank you

[crystal mortar]

Hello. So i have two issues with Suno that i would like to adress.

First is about continuating from same song. I do that when i see that song is good, but some part of it are bad, like missing words, or singing with wrong articulation. So i start over from certain place.... only to realize that it recreates mistakes as well. Even if i change the prompt to give new direction. So that makes impossible to "fix" the song version you like.

Second is about generation with misplaced lyrics. I am getting A LOT of that, sometimes the entire song is bugged and not a single line in it's place. I feel like those kind of generations should not occur at all, or at least not waste credits. I think it's fairly easy to set a check that will see of AI follows the lyrics from the prompt, and unfair to charge users for generations that are bugged.

Also would be nice to allow re-creating certain part of the song without touching others. Even if it's first part.

Finally, it would be nice to have formatting tags that AI can't ignore. Because RN i have a lot of tags ignored most of the time, even simple ones like [Short Instrumental Intro]

[harsh musk]

Up

[opaque cipher]

Hi Team, just wonder at the problems with inaccurate pronunciation at the end of a Chinese song I've composed (most of it will be fine at the beginning), can this be controlled by tags?

[harsh lichen]

hi, i have a issue with suno, i liked a song yesterday and wanted to listen to it again. i had asumed that you could find it easy after you liked it by searching the name or going to your liked songs but it turns out tjat these are not features. if been looking for it for a while and can't find it anywhere. i would like it if suno had a feature where you could find your liked songs by going to a liked section or a search bar where you can find the song.

[scenic girder]

hello,I would like to put forward a functional suggestion, that is, can the music be divided into tracks, divided into vocals, background vocals, and each instrument, and then for the user to choose and combine? I need this so much. Hope it gets adopted. Thank you guys!

[tidal wharf]

short answer: not really, or at least not easily

long answer: the way the AI model works, it generates the entire master track at once, as it was trained on complete tracks. To train a new model means you need to have original divided tracks for thousands, perhaps tens of thousands of tracks.

the best the team would be able to do is to apply some splitting post-factum, but there are already tools for that. (I use vocalremover.org but I don't know if it's the best one.)

[narrow knoll]

The Suno App needs a way to see the User names of those that create songs, as well as a way to contact them if they choose to display that. Is there any plans on that?

[astral shard]

make plan for unlimited generations

10$ a month for 2500 not worth it when 1500 are free

[solar socket]

You can download the song then plug it into Moises which is an app

[hybrid crane]

Depends.. if you are loyal and come back every day, yes.. you have 1500 credits. However, it will not build up. So if you come back 3 days, later it is still 50 credits and not 150 credits. And unlimited is just not feasible. GPUs are expensive.

[hollow locust]

For lyrics to never start before around 00:10. It's such a nuisance when they start on 00:00. And why does 'sultry female vocals' trigger it to start so soon?

[loud canyon]

I think the more important differentiator between free and paid is whether you own the rights to the song (you own the rights if you pay) For some, they use this as a tool for music ideas so $10 is worth it. If just making music to listen to for fun, you're right, not much different

[warped finch]

the extension (continue song) must be 2 mins

[novel compass]

The ability to edit the lyrics after the song foundation is laid so you can make small tweaks to make things fit more tightly.

Much like how real songwriting works I suppose.

[shy lance]

i would also love to have "seperated namespace" for creation mode. so all songs will be automatically grouped into a certain playlist for example

its really hard to go through a list of thousands generated songs mixed up because you work on different once at different times

[novel compass]

Yeah, the whole UI could use an overhaul, IMHO.

[spiral marsh]

a way to edit the lyrics to follow the generated song? Especially where a song is continued where part of the text is repeated in continuation to be able to complete whole song.

[spiral marsh]

A way to like a song as "I will buy that" in stead of "oh that is an interesting mess" 🤣😂

[quartz bronze]

Make it possible to generate 1 at a time and not instantly 2

[raw knoll]

Showing how long a generation is without having to click on it next to the thumbs up/down.

[plain sedge]

A way to extract the thumbnail image it generates, so we can use it for Spotify/Soundcloud etc

[limpid moat]

Something i might suggest as if things are spelt incorrectly it will still try to sing them that way, not saying that you spelt things incorrectly but purposely spelling things badly might get you the sound you want, i had a similar issue with Moonlight, i just added a space in the word and it was sang/spoken better

[sonic laurel]

how about incremental updates instead of having to wait for a whole has new version?

[wooden fable]

modern webdesign consists mostly of empty space!

the more space, the more modern!

[hybrid crane]

this is why:

It's not dead space if you have a long "style of music"

[upbeat sage]

Suggestion :
Embeded playlist like spotify on Discord

[hybrid crane]

I agree on that. I personally made some proposals too, but small team.. million things to do. They are working on a better experience.

[wooden fable]

i have a feeling the devs are swamped by the sudden popularity of this site, so stuff like this might not be addressed right away.

how to you prepare for a tsunami? 😅

RUN !!!

[hybrid crane]

GPUs.. GPUs... panic.. more GPUsss

[crystal river]

i cant do a new acount

[hybrid crane]

I don't know why.. can you give more information?

[crystal river]

Bot traffic was detected. Please try again or contact support.

I get this

i dont know why

[hybrid crane]

we see a lot of Spanish users struggling with thise.. but no idea why (yet). A temporary solution is to use a VPN. If you don't have a VPN, I can suggest using https://protonvpn.com/free-vpn. It's free and safe to use.

[visual brook]

could you ask the staff to provide a prompt describing their operational and data-science challenges (without giving away proprietary information)? would love a poetry song from the engineers in whatever style they please

"Increase instance size, scale DB, volume projection for black friday, retrain models on linear sub-quadratic context length audio transformer model" etc

[crystal river]

i do all but dont happend nothing

[hybrid crane]

The only think I can think of, is resetting your modem.

[hybrid crane]

I like your humor 😉

[languid sequoia]

Is there anything on the roadmap for vocal variety?

I keep trying to make songs with vocals like Louis Armstrong, Tom Waits, Thurl Ravenscroft, Cab Calloway, Leon Redbone, etc... and i just get Michael Buble, Dean Martin, or Travis Tritt...

[raw cave]

https://tenor.com/view/spinning-prize-wheel-wheel-of-luck-lucky-color-wheel-gif-14957542

[languid sequoia]

haha - that's exactly what it seems like... even with bracket tags in the lyric section...

[lapis osprey]

Need more length in custom lyrics... for faster songs it runs through the limit quickly, forcing to extend the song where not needed.

[warm zephyr]

anyone find a way to fix the unwanted vocal harmonies?

[raw cave]

Yes, re-gen.

[wooden fable]

that usually doesn't work though

you pretty much always get them these days

[raw cave]

You're correct, I should have said re-gen, re-gen, re-gen, etc.

[wooden fable]

no wonder the servers are overloaded!

[floral moss]

if or when we can have the original prompter added to posts, we could have awards that in turn transfer credits like a thank you to the prompters. so a free user could even get enough credits to keep generating if they're on to something.

[limpid moat]

So vexed right now 10 reattempts on a finally got the perfect one then suddenly the last 2-3 seconds start up a whole other beat, I know I can splice it manually after download but a Splicing feature would be a dream come true 😛

[old relic]

What is the motivation for Suno to reward free users?

[woeful rampart]

"Similar Songs" for exploring other peoples prompts please

[floral moss]

a social credit system would be rewarding quality song prompters, this could encourage more new users, encourage users to share their generated work, and sharing some credits from those that can afford them may increase the number of credits sold

[obtuse rampart]

wich tells me this

[gritty dock]

A way to search through personal songs and sort by date and genres and names as well as a way to do something similar of others songs

[lament summit]

I would like them to add the option for seed/random.

[tropic slate]

would be nice to have the instrumental and vocals generated separately so that if you only like the beat or vocals you can keep working with the one you like whilst fine tuning the other to work better (mostly lyrics cuz sometimes you get the perfect beat, perfect vocal, but some of the lyrics just don't sound as good as you'd hoped and wanna edit them a bit).

[tight beacon]

Usability Interface: Let users mark clips for deletion with a checkbox, then have a button that sends them all to the trash. Confirmation Dialog "Send all clips to Trash?" should have two buttons "No" and "Yes, I'm sure".

Same for Trash except it should have "Delete Permanently" Confirmation Dialog: "Permanently delete these clips?" buttons: "No" and "Yes, I'm sure".

[tropic slate]

no idea if it's possible, but just would be a neat quality of life addition instead of having them merged as one right off the bat 🙂

[wooden fable]

if i could get the vocals out by themselves, i'd create the music in a DAW

but using a vocal separator destroyes the quality too much IMHO

[tropic slate]

I just want the ability to train on the same song to keep everything the same bar like a word or two

similar to what you do with AI art or ChatGPT prompts

[marble gyro]

they need to add seprate back up and main vocals - kits reads it as 1 fulle file

[tropic slate]

just feels bad when you get something so fucking good but then realise there's an accidental typo in there or you think of a better line for a spot

[hushed stump]

Suggestion for basic plan users:
Add an option to purchase songs they like for a set price.

For example:
If users create a song before subscribing to the Pro plan and really like it, they might be unable to recreate it once they've subscribed. This option would provide them with the perfect opportunity to own those songs, alongside serving as a suitable alternative for those who don't have a set monthly extra income.

Additionally, this feature could generate extra revenue from casual users who are trying out Suno but don't necessarily want to continue using it. They may, however, appreciate the idea of owning one song.

This addition wouldn't alter the basic plan; it would simply introduce this purchasing option.

Considering the price based on the $10 monthly billing for the Pro plan:
1 month, $10 ÷ 500 songs = $0.02 per song. Therefore, the price per song could be set higher, somewhere between 10-50 cents.

[tropic slate]

I'd be down for that ngl

would be real nice for us who are doing it as a side hobby as something to fill our days with when bored

[grave pecan]

Please, for the love of AI, give us negative prompting. I'm tired of vocaloid vocals slipping into my gens.

[worn atlas]

To unlock all the potential of suno, it should be possible one day to have fine-tuning of the model using user's own data. The user can collect at his own, musics in a specific style and train the model with. That will spécialize suno in a particular style. Fine-tuning models becomes common in AI world.

[eager inlet]

I would like to pay 10 Bucks for 1 Month Premium Acc with Paypal. I only see MC etc. which I cant use.

[vast steeple]

few things I would like Suno to add:

1. Ability to download with higher quality (studio level).
2. more instruments support.
3. ability to implant my own voice

[grave pecan]

For 3 you could split out vocals and then replace them with your own voice using RVC

[tropic slate]

rvc?

[grave pecan]

https://www.youtube.com/watch?v=aVfoC1EHlVs

Just search for RVC

[viscid cloak]

A Status at the Side of the Website: (Like Green, everything is perfect you can generate, Orange it takes some time, Generating is disabled)

[hearty star]

It would be great if songs listed their author with a link to a list of their public songs; currently there is no way to know what other cool songs a user has produced. Also I personally would like the ability to crop the song beginning or end. Sometimes I want to clip 20 seconds off the beginning and 5 off the tail before I publish it ^^.

[steady scarab]

today udio, in a few more days another one coming, please fix this asap we love you guys:)

[static swallow]

Anyone know of a service that would let you upload a tune/melody and have it create an arrangement based on that? Talk about a great way to do demos!

[delicate comet]

Ability to upload own instrumental for vocals and also download in wav at a much higher bitrate

[jovial silo]

We need a real-time status channel on Discord where we can see exactly what's happening and how long things like maintenance might take, for example.

[hybrid crane]

This is pretty real time:

And end time is unknown

[jovial silo]

I think stating it here is okay for the website, but this way, it's easier to follow and understand.

[hybrid crane]

Aber doch nicht auf Deutsch lol

[jovial silo]

Kann man auf Englisch

[hybrid crane]

Ja genau 😉

[jovial silo]

Ist ja nur ein Beispiel 😄

[jovial silo]

But actually, such upgrade tasks are quick for me; they're done within 2 hours, lol.

[prime rock]

I know the place is a dumpster fire right now but could some mods actually do something about the people in chat who are outright harassing others?

[languid canyon]

Make suno paid only if you cant handle the infrastructure . Have a landing page with a short introductory video about a succesfull workfflow preferably from a trending song . Show how one can make a song in a 1-2mins video. The fame is already out there and people are aware.

[worn atlas]

There is a degradation in the instrumental quality after the Split. Isnt it?

[grave pecan]

Depends on what model you use, but sure none of this is technically lossless. I have found it to be good enough

[raven vale]

If this hasn't been suggested maybe, a way for Suno to know when its an end of the song.

[grave pecan]

[end song]

[wintry moth]

Man I came here for that exact issue because i had a song with literally the PERFECT BEAT and it starts off the song with a large harmony

🥲

[jovial silo]

Macht doch mal eine Ankündigung vllt. bleiben dann die Leute hier lockerer :3

[hybrid crane]

I am Dutch.. I know German a little. And I am only a volunteer. But I will let them know 😉

[jovial silo]

❤️ Bedankt

:=)

[still breach]

That works? lol

[potent folio]

you do too much for a volunteer 3daisy, you have my respect 🙏

[hybrid crane]

Trying to keep people updated and calm.. thank you for noticing. 😄

[still breach]

If it hasn't been suggested yet how bout a trash all dislicked option or automatically trash disliked clips

[sharp violet]

Pleas let me make longer songs from the start , the 2 minute limit is insane

[worn atlas]

Find a way to add chords in the lyrics

Something like that
[Verse 1]
{C#m} This is my {D} best song
{G} For Suno ...

The chord is written just before a word of the lyrics

[hearty star]

I just checked my songs and it shows I have no playlists and no songs... what's up with that?

[fossil kindle]

Hey is there a way for me to link my email? I already have the discord, but I want an extra recovery option.

[acoustic pike]

Hello admin How are you?
Please Add 'Solo' and 'Duet' Singers option in V3 or V4 like V2. Male and Female voice word not working on there.
Please try to make it that Male for "M" and Female for "F".
V3 always make Group singers voice like Chorus Singers or auto-tune sound.
V2 made good voice quality but there is not full song.
Hope you understand what we facing the problem.

[paper maple]

Not sure if it's been mentioned but a way to use the same "seed" if there's need to fix things like a lyric could help. Also line breaks to ensure it doesn't rush into the next line would be good

[tawdry scroll]

💰 Don't charge 10 credits if it generates 3 seconds of noise. -> credits linked to generation length

[tepid trench]

Anything under 8 seconds gets automatically refunded (though should be a bit longer imo)

[grave pecan]

Honestly, they should prorate credits proportional to the length that is returned out of 2mins

[acoustic pike]

Hakuna Matata
2 male voice for the song
M1 for Timon
M2 for Pumba

If i want to add a kid voice for Male...
MK1 as Simba.

[tawdry scroll]

I have so many places where it generates 27 seconds, that should cost 2.5 credits not 10 in my book

[acoustic pike]

M for Male & K for Kid

1 for 1

F for Female K for Kid
1 for 1

"MTV" for Male Teenager Voice
"FTV" for Female Teenager Voice
"MOV" for Male Older Voice
"FOV" for Female Older Voice

Mtv sounds like music tv
Ftv sounds like fashion tv
lOlzZ

Anyways..

[tiny cloak]

Has been said but
Anti prompt
larger style box,
Dominating instrument
bpm,
rhythm,
sample taker or smthng to draw inspiration from if you want to create a similar feeling,
changing lyrics in post production of the same clip
and being able to add intonations or curated how it some words should be said rather than generally the whole verse,
pitch modulator,
voice and instrumental separator,
being able to take a created part and then muting the vocals,
set vocals throughout the parts

[acoustic pike]

M1, M2, M3 for Male singers
F1, F2, F3 for Female Singers.

Chorus Singers for "Chorus"

I think
Male Chorus Singer (MCS)
And Female Chorus Singer (FCS) add there also.

[kindred sonnet]

Adding a function of karaoke by removing human voice and following text sub title as well as singing guide in video screen.

Adding a function of replacing desired voice cover would ve perfected the Suno AI Music Service.

Ability to choose voices

[jovial silo]

Dirket Helper Bot auto Infos when asks.

[halcyon escarp]

Small but very useful feature:
The ability to download all songs

Right now you can't download a song unless you do it from the trending or create pages
If I pull the song up in a new tab, the option to download it isn't there (unless I'm missing it and I'm just an idiot lol)

[sage shore]

it was there earlier, they're doing site updates so it was removed for now it seems

[pallid notch]

can we have a mobile app? using website on mobile is not ideal

[ashen stirrup]

separate cesspool for funny/dark/horrible/offensive songs

[vapid isle]

comment section for each song, people can talk about the songs or rate it with own words

[ionic skiff]

nooo

[formal dove]

Is it just me who does this? Is it unusable because everyone is updating?

[hybrid crane]

Everyone.. they are moving databases and it will take a few hours: see #💬┃general-chat message

[misty solar]

A way to delete generations we don’t like, so I don’t have 40 pages and it’s easier to find the ones we liked.

[torn lantern]

my number one would be able to train my voice to suno and have it use that specific voice (or whatever other voices I trained) in the song... are there plans for that? is there a public roadmap to see where suno is headed? thx!

[hybrid crane]

Roadmap is not public. And training voice is interesting, but has many hurdels like copyright. How to make sure you are training your voice and not a celeb.. 😉

[median tinsel]

Suggestion: 2$ plans with 100 songs per month

[dusty fable]

It would be nice to be able to have the option to download the music track separated from the verbal track so we can use a DAW. I had a few good songs that sounded great with awesome headphones, but crapy everywhere else, like the phone mono speaker, or cheap headphones, I couldn't understand the words with the cheaper headsets or off the phones.

[torn lantern]

true - or if there is a way to download the music and the voice as different tracks then? could then use some other tool to take the vocal and transpose into your voice (i think some tools like that exist already but could be wrong)

[amber finch]

i've been messing with suno. it would be awesome if there was a quality "upscaler" for audio tho, some of the gens are great but get buzzy or have audio artefacts

[honest coral]

I love Suno, blew me away today, first time user. I just have one issue: The artist name filter seems insanely restrictive, I can't use many of my friends names and my suggestion is to make the artist filter take context into account, as of now it seems it is a simple keyword filter. That does ruin quite some usecases. One example is I wanted to make a song about the 101st Airborne, clearly not an artist and it refused to do so because apparently "Airborne" is an Artist.

[neat belfry]

Hi developers, any idea whether we can download the intrumental version of my song? My friends like my songs very much and they want to remove the vocal to sing them. Thanks!

[dusty fable]

Also editing a created track is almost a requirement. For example, I used a song that I wrote, but had a typo in it. The track was almost perfect except for that one word used in a few paragaphs. It would have been nice to edit that mistake out, After several wasted credits trying to get it to sing it in the same way previously, I gave up on that. So many frustrations and wasted credits, it says 500 songs but Doubt I'll get 10 (20 if I stop trying to tweak every mistake).

[glossy cape]

A page or list of recognized tags, instruments, vibes, etc. would be cool.

[formal bear]

we have some but not super great yet, will give it to you when it is in better shape

[dawn fjord]

Multi select or playlist downloads.

[amber finch]

niiiiice

[formal bear]

you can already

press control and select...

[dawn fjord]

nope. download disappears w mult select

[formal bear]

oh you can add to play list I think

download can't

[dawn fjord]

yeah thats all, or trash. nd you cant download playlist - so only way is a laborious 1 by 1 menu download

[formal bear]

yeah...

k

should be very easy though...

[dawn fjord]

yep seems like an easy fix that would be v helpful

[neat plank]

One of the ideas I can propose is that once the song is finished, you can edit the lyrics. even if it costs 10-20 coins per song (I was able to create a perfect song with an excellent soundtrack, but due to poor spelling it sounds bad)

[prime rock]

Not being able to get to half the options from the sidebar on the create list is really obnoxious, because the three dots in the main window don’t respond at all half the time and it’s difficult to keep track of what you were doing

[median tinsel]

What should have been implemented in v3 is the acapella voice without chorus or harmonies

If you adjust this (options: instruments, unique voice in the song, etc.) it will be even better.

[honest coral]

Is there a way to force a vibe change during a track? For example have half the song a calm cinematic score transitioning into a fast percussive beat, shifting the vibe from a foreboding to a "chase" one or something of the like?

[vagrant coral]

We're all adults here.

Add a opt in prompt for
"Couldn't generate that.
Prompt flagged for moderation"

[lyric meadow]

I am from Turkey. Suno is a very good Ai. I want to buy it but i can't because pricing is too much for my country. Can you make special pricing for turkey ?

[amber finch]

any plans for audio-inpainting? selecting a part of a good song that could use touching up or redoing?

[feral aspen]

Supposedly being worked on

[amber finch]

very nice

[mint burrow]

Please allow authors to change the lyrical text for their song 🙂

[solar yew]

Make one part of a song, e.g mellow hip hop. Select 'continue from this song' from the options menu to make a "part 2" of the song. It allows you to write more lyrics, do a style change, and even choose when the change occurs in the original song.

Suggest making the timecodes more specific when it comes to doing a continuation? In trying to do a continuation for a certain point in my song, the continuation happens like a half-beat early or half-beat late.

[peak pasture]

Suggestion: Refund Extended tracks under 1 Minute, because the most are 14 sec. and trash!

[solar yew]

Messes with the flow.

[still sky]

May be a big ask but if there’s any way in the future of the project to manually post edit the music that’d be awesome. Like a very very basic FL studio in the site or a file that opens in another program etc.

Sometimes instrumentals can be good but if the 808, hi hats, etc are just slightly off the whole thing goes down the drain

[limpid moat]

On the Side bar can you add a folder/playlists button for adding completed songs from your library. Considering there is no Folder option I just use a playlist and add my completed songs there.
Example of Request

Explore
Create
Library

Folder
Playlists

[leaden marlin]

I turned on instrumental to make instrumental music. I clicked to turn it off. Then the next 2 songs were instrumentals too even tho I turned off instrumental

[upbeat rock]

“Lyric” or “Instrumental” tags for at-a-glance info, labeling, and filtering library

Place these tags here

Maybe find a place for username beneath the art thumbnail

Created by: [user] or simply link to [user]

And obviously profiles will be a gamechanger

Someway to nest extensions to the base track will be extremely useful for better workflow

[violet crest]

#💡| Have a « Default Mode » we can choose between Instrumental and Lyrics song. I am diving in instrumental and how many credits I lost just because I forgot to click on instrumental.

[feral aspen]

You can also create instrumentals without the toggle, I typically prefer that actually, so I can use lyrics to try and guide the instrumental like [rising strings] etc. (You may sometimes need to add "instrumental" to the style prompt to stop it from singing random words, but that's usually all you need to stop that.)

[dusk rune]

First of all, thanks for making this AWESOME. I love it even more as generative art.

Mass deletion would be appreciated.

Also, adding to your YouTube channel as a short would be a blast.

Maybe test by producing one instead of two. I usually like the first one better, but I could be wrong.

If it is a male or female voice as a toggle button, I know you can add it in the prompt but some people like buttons.

Add staff-picked songs.

[lyric meadow]

@feral aspen Have you plan like that ?

[feral aspen]

As far I know there isn't a plan to have country specific pricing, but it is certainly a feature idea you can suggest here.

[peak pasture]

Plz remove the loud backround noises from Eletronic, DNB, Dubstep,etc. Tracks.❤️

[sage mirage]

The UI imporvements are grand, thanks guys. Especially like the change in labels since it follows logic a lot better. Plus the 'Extend' button is A++

[upbeat rock]

Been there

[peak pasture]

Suggestions [ Overall Quality] : Remove the squishy Backround Noises (that sounds like you sitting in the Bathroom) in following Genres: Dubstep, DNB, Electronic,Melodic, Harmonic and Classic- Music. [Refund]: Plz Refund Extended Songs shorter than 1 Minute, because the most are 14 sec. and not worth for spending Credits on (example video). [Timestamp]: add a Timestamp and Date to all Songs, so we can see the Date & Time of Creation.

[tawdry scroll]

try RipX Daw, it splits your track into layers which you can then edit individually. You just need to copy the correct word from one suno track to the other.

Until suno exports individual layers, this is a good workaround for me.

[dreamy arch]

I need a rapper with a lisp

[flat portal]

Hi guys, can someone tell me how best to incorporate the start and intro into the song? Because the song starts in the middle and I want to build the song slowly from the melody, I don't understand how that's supposed to work. Kind regards

[vagrant rock]

Please include 'continue from' and from what second with a 'reuse prompt'. I was working on something yesterday and now I have to find those settings again 😦

Erm, where did the 'continue from' go? This option seems to be missing?

[ionic jetty]

HI , can you implement a search functionality with filters into the web interface. The library gets large and its not easy finding and revisiting previous tracks especially if you deleted most tracks or regenerated them.

[mint thorn]

Damn please unlock the censor on lyrics section... 😢 it’s such a blocker to creativity, is better to create a list of of forbidden words ?

[woeful rampart]

Add a History to the Explore page so I can see what songs I listened to before

[leaden marlin]

I think it'd be cool (Of course not as it is not, as it would take far too long to generate) but discectable sound files so it generates vocals and instruments seperately allowing for Vocal modification later

[rancid marsh]

I still wish stems for a paid account.

[hazy fjord]

Hey, dear developers, could you add a Hardbass to Suno?

https://youtu.be/BnTW6fZz-1E?si=ZoZakQkC9n8edHzd

[supple cloud]

I would like som sort of folder system, like Favorites, but you move the song ( not to trash thats to messy), not just link it so I can sort out Full songs and so on

I would like a "Get Whole Song" option even on the first tracks produced because when you create a short song it won't have anything to continue with but i would like to be able to tag it as "Full song"

[feral aspen]

As a workaround you can use V2 to continue and set style to "end" and prompt just "[end]" in lyrics, and it often just ends within a few seconds. And then the song can be considered a Full Song.

[final cairn]

I suggest to remove all censorship but disable public / direct links to a file that has inappropriate content allowing only for direct download

[supple cloud]

OK, i'll try that thanks. by the way I love your new updates today

[feral aspen]

That's all Suno, I'm just using their stuff and helping folks!

[supple cloud]

lol

[full sky]

so where is the suno music discord bot?

[soft meteor]

The new Explore page looks really good.
But it could benefit from a searchbar for the genres currently trying to listen to specific genre samples is as fruitful as clicking the random button 😄

[warm solstice]

Some suggestions
• The site needs a proper Search Function. I should be able to search others' songs by title, lyrics, and genre/style in a proper gallery-like format (similar to Midjourney's website).
• Not sure if this is possible yet, but in the future, I would like to see the ability to turn my own humming into melodies to better direct the flow of what I'm generating instead of relying on pure RNG to give me a catchy melody.

Regarding Explore Beta specifically
• Assuming rating songs (thumbs up or thumbs down) is used to improve the quality of the model, the Explore feature should have these buttons visible.
• There should also be a button to easily copy the prompt of a song playing in Explore directly into Create.
• The volume control button in Explore does not seem to work, I can only Mute/Unmute but can't adjust it otherwise.

[prisma ridge]

The Volume can be adjustable too, not only mute or max

[tight wave]

Fix the ability to use it in DMs with the app.

[karmic snow]

Pro subscription (The best option) lets you generate infinitely without the worry of running out of credits which also gives more of an incentive to pay the highest cost

[warm solstice]

+1 for the feature to mass delete songs. It can be a bit of a pain to keep my library clean when I have to manually delete dozens of song extensions that are just cluttering up the place.

[karmic snow]

just a simple checkmark box and a trash can icon addition would solve that
EDIT: or just putting the "move to trash" button next to the "Copy Song Link" button for faster deletion

[autumn acorn]

A way to edit the exact same track. Just minor edits like word spellings, which it should re-render, while preserving the remaining music

Why ? I absolutely loved a track Suno created, but it got the pronunciation wrong on a few hinglish (hindi written in english) words, which could be improved with some simple spelling changes. Plus the lyrics itself had mistakes. I want a way to make changes to that in the track, while keeping that rendition and everything else, the exact same.

Is this possible ?

[woeful rampart]

"Shuffle All" button on the Playlists page so i can get songs from every playlist i have saved

[loud canyon]

Filtering of library > liked. I would like to see songs I've made that I like that are part 1 or whole song. (not part 2,3,4 etc.)

[empty geyser]

There is still one, and it is the most expensive in its development, known as the surgery studio for music generated by Suno. If this exists - FL, others can go to the next world.

[outer niche]

Allow to regenerate same song, at the very same style (fix seed) and change the lyrics a little.

[karmic snow]

Ability to "dissect" a song. Doing so tells you more about what instruments, instrument style, vocal style, tempo, etc and whatever else that can be used in a generation that it wasnt directed to generate

[violet niche]

Allow users to leave comments on songs

[woeful rampart]

When Extending a song to overwrite old lyrics, when the you press "Get Whole Song" cut out the lyrics that were overwritten

[dry robin]

can we sort the guitar sound out, i cant get a full good song the mix is horrible and vocals are weaker than they used to be, whatever happened 8 days ago has really messed up being able to write rock or metal songs i have about 20 songs that sound aweful

[peak prism]

First of all, I wish SUNO continued success. Please enrich the variety of Chinese instruments in your instrument library. After trying dozens of times, I've found the guzheng appearing consistently, its jangling sounds becoming quite intrusive. The bamboo flute surfaces only on rare occasions, while stringed instruments have yet to make a single appearance. I don't believe this issue lies with my choice of prompt words. I suggest you enhance your instrument collection and develop methods to produce genuinely authentic Chinese ancient-style songs as well as popular tunes. With a population of 1.4 billion, there is ample opportunity to earn money; all it takes is delivering high-quality music. Keep up the good work!

[primal pawn]

The bass sound is always too loud after making it.

[visual brook]

You may want to try more prompts, or start a side project with other Chinese to try to find the best prompt style tags for your goal

The model likely has what you need, but the prompts will be very challenging to find, you could try "traditional Chinese" "folk Chinese" "{province_name} folk Chinese" and other prompts

It may take some time to find it, and not sure of the prompts you tried

[lapis osprey]

My biggiest issue with generated songs is the clarity. So often the voices (and instruments) sound muddy instead of clear. Seems like, frequently, the longer the song is the more muddy they become. Also if a song starts off clear and becomes muddy sounding by the end, if you extend it, the extension retains that muddiness instead of the clarity at the start.

[earnest nimbus]

Paypal for Subscribing would be nice

[tidal isle]

I think the most important improvement right now would be to be able to re-pronounce words that were mispronounced, but keep everything else exactly the same

[dense monolith]

A cool feature would be to generate lyrics without generating vocal voices... so you can still can sing over the song and let ai generate the Song itself because... the lyrics are awesome in Suno!

[feral aspen]

Well you can generate instrumentals (with the button or with style "instrumental") but I think mean you type out lyrics, but then Suno does not actually sing the lyrics right? That would be cool feature. I have tried to do this, I called it "lyrics guided instrumentals" and it is not impossible but also not very practical except with lots of luck.

[dense monolith]

Other way arround you generate instrumentals but lyric text will be generate even without vocals so you are able to sing them yourself

[feral aspen]

Oh interesting. That would also be tricky, because the lyrics would have to match the rhythm of the song, the words have to fit. Cool idea.

[autumn acorn]

Mark which playlists I added a song to already, if it has been

[quaint path]

hi folks. thank you for making this app. it is the best app i have ever used by far. my suggestion is to allow for a larger scale/spectrum of liking a song that is produced. i want to be able to "super" like songs, which would distinguish it from just a regular like. some sort of gradations would be helpful. think super-like feature on Tinder, except without limits. this feature would be so helpful in helping me sort the true gems out from just the regular 'likes', cuz im generating so many songs. thank you for your this consideration. love the app. you people are the best humans

[urban sleet]

Once again - I soooooo hate this button being turned on everytime the page reloads... 😡

[quaint path]

this exactly. good point!!! i always add stuff more than once cuz i generate so many songs

[upbeat rock]

💡 Please consider some updates to this server. There's opportunities to expand it in deliberate meaningful ways. I've done my fair share of server design. Keep it simple but add functionality. DO - section for serious music vs joke music. DONT - channel for sharing pictures of pets. Keep it focused on music and give people the place to collaborate and build community

[upbeat rock]

you can do something kinda of similar by filling the lyrics box with only meta tags

[upbeat rock]

probably the most requested feature. A true remix option. I am hopeful that the renaming of remix to reuse will mean this is coming sooner than later

[kindred aurora]

Pls. Add volume slider for tracks.
Whare track start, he start on full volume. Its soooo pain for me

[carmine grotto]

Make a mock-talking feature for diss tracks

[blazing sinew]

I think a tempo slider or something would be cool. So you could decide how fast or slow you want the song to be (Not necessarily talking about the length of the song)

[sturdy ocean]

Here are a few suggestions:

1. Create the website -- status.suno.ai -- that is updated every time there is a service disruption, or system change so we can subscribe to the updates and know what's happening with the service.

2. More dev transparency in posts. There are still generation problems that are unsolved and unaddressed. No gaslighting allowed! We need direct answers from those in power as to what is happening and why and the expected results and fixes on a timeline. To still offer V2 with V3, but not the V3 Alpha is something that doesn't' make sense. Explain to us why.

3. Create a private conversation area here for Pro users only. Most other Discord servers offer that sort of private place for paid conversation, insight and discussion with access to the paid staff for direct inquiry.

Thank you for reading this!

[potent rain]

Definitely need #3. General is a mess

[sturdy ocean]

Yes, it's curious they don't have a private place for those who pay. Sometimes there are like 5 private text channels for Pro users to help break up the overwhelming flow at times.

[heady sluice]

SUNO AI definitely needs some type of editor. When you try to create a long song, the final complete song always cuts pieces of the song. It’s pretty annoying because it cuts part of the lyrics.

[potent rain]

They dont have many devs and it takes time to plan out and implement authentication between discord and their user registration so it makes sense why its not implemented (the bot also fell way behind with v2/ v3) but its definitely annoying

[sturdy ocean]

There are Mods here who could be empowered to set up the private areas and moderate them. The conversation there is usually much less heated and more intuitive because the Pro payers have a lot more experience with the system. Then the Devs could update the Mods if they didn't want to get into it or feel targeted.

[tawdry scroll]

see here #💡┃suggestions-web message

This channel should be switched to the tread / forum style discord like #1101325881059840040

[heady sluice]

Since I can’t memorize, I used SUNO AI to help me with the process of memorization. I uploaded my entire presentation, but it always cuts words and doesn’t follow the same sequence or seed.

[potent rain]

At the very least it means making a backend portal for mods so they can verify paying users, exposing privacy issues, giving more access to volunteers (which afaik they have zero of atm), etc.

[sturdy ocean]

Well, we login to the website with our Discord credentials, so they know who pays and who doesn't pay. We'd probably have to leave this server and rejoin to get the "paid" flag set, but that's not hard to do. Many Discord communities us that method for automatic verification and role application.

[tawdry scroll]

its easy to automate this assigning roles to subscribers, I've done the same on my discord server

[potent rain]

You do, but theres two other forms of login

[tawdry scroll]

they already use clerk so it should be just configuration

[potent rain]

“Should just be” - as a software engineer those are famous last words

That said i dont have experience with managing discord servers

[tawdry scroll]

hence *should

I do the same on my discord server, its kinda trivial to setup

[potent rain]

With all SSO? Or just the integrated discord SSO users?

[tawdry scroll]

SSO

[potent rain]

What?

All SSO or just Discords?

[tawdry scroll]

all that clerk provides

[potent rain]

How do you know theyre using clerk

[tawdry scroll]

looked at the api calls

[potent rain]

Ah its some auth thing like okta. Thought it was discord management or something

[tawdry scroll]

suno uses clerk, discord uses ...discord

[potent rain]

As in third party management for discord servers, not discords own thing

[grim cedar]

You can also login via google accounts, in which case they wouldn't know

[frosty ferry]

Heres a suggestion that Id like to see implemented for Suno, in the create - Custom mode, Id like the option to use a AI generated picture, OR have the option to use something that we have locally on our PCs. That way, all the clips that we generate will consistently have the same image if we so choose.

[reef marten]

I have these problems since this passed weekend that I didn't have before:

1. continue seems so random it's almost useless
2. frequent glitching out on lyrics after a continue
3. lyrics have lost any adherence to chorus/verse structure

can we have the old "how it was two weeks ago" as a separate version # on the drop down. call that 3 beta or v3b or something ?

and can dev please stop altering the version of the code that the users are using? Make a stable release, have that be a consistent piece of software that doesn't just suddenly change each weekend ?

[tiny dragon]

If there isn't already make a way to add "comments" to how the song should change for example "* louder *" since it only rarely works and sometimes even sings the text

[ruby estuary]

I would love to be able to generate two different style of the same track, opening possibilities for "breaks"

[empty geyser]

loudering

[pallid notch]

api for developer?

[limpid moat]

Try using the externd feature and on the left where it has Style of music, keep what you have but also add louder
eg.
Synthwave Pop Electro LOUDER

Also untested hence me saying try, adding things to the style has so far worked wonders for me, but remember this is a baseline generator, you are supposed to DL your full song and run it through a mixer for things like changing levels 🙂

[tiny dragon]

I'll try that, thanks :)

[peak pasture]

Suggestion: [Commercial use] : Marking of songs that may be used commercially and those that are only used privately would be fine.👍

[rain ore]

Undelete or at least an ability to access my trash. I got misaligned on a click and deleted a song I wanted to keep

[limpid moat]

yeah i would like to know that if you extend a song you made while free but are now extending it as a Pro account does it then become commercial use as a full song after multiple extensions to form a full song.

[fringe cobalt]

It would be cool if we could upload our own tracks and type in lyrics and the ai generates vocals for our tracks

[limpid moat]

I imagine that stuff is on the way 😛

[fringe cobalt]

I hope so, I'm new here do the devs read these suggestions?

[still breach]

Do paid members not get daily credits? I feel like we don't or atleast I'm not getting any I'd definitely suggest that we should get atleast some daily renewal

[limpid moat]

yeah last night there was server issues and most stuff getting suggested got added after the fix simple quality of life stuff but this is blowing up at the moment so they busy bees.

[limpid moat]

nods 3 days in 1280 credits left

[still breach]

3 days in I have 350 after dropping an extra 4 bucks

[fringe cobalt]

I can imagine, it has a lot of potential I won't use it for anything serious until I can do that but I'm still having a lot of fun with it at the moment

[paper maple]

The credits-to-paid account feels a bit off balanced. $30 gets you a month with 10k...but $30 topping up only gets you 4k?

[still breach]

Never noticed that lol but you right

[paper maple]

I could pay the same $30 to make a second account and get 2.5x what the top up offers.

[still breach]

Very true

Maybe lower the cost of non custom generation to 5 credits

[fervent spade]

When you release a updated model I would hope you would have trained it on more modern music, especially like Phonk music, Brazilian Phonk, even some save Europe stuff

[limpid moat]

Expand on the Style of Music section in Custom, add all the terms associated with the type of music you want. Even put the countries name, like if you want Brazilian Phonk don't type it like that, type Brazilian-Phonk

that will narrow the learning

[fervent spade]

Yo it worked thx 🙂

[limpid moat]

Your welcome buddy 🙂 been wanting a haunting female vocals so i had to go with,
Style of Music
female-vocals melancholy synthwave haunting eery eeire sad fearful

[fervent spade]

dang yeah I've realized to get the specific thing you write it in the style and I even put it in the brackets in the lyrics like [Emo rock Guitar Transition]

[limpid moat]

I want a way to add noises, like scream birdcall eagle screech seagull DogBark Carhorn stuff like that 😛

[peak pasture]

Yes, this tempts customers to delete their subscription and make a new one, which cannot be in the interest of the operators

[fervent spade]

yeah they're subscription model is lowkey weird hopefully they fix it especially with that new Udio AI music thing out now. They got competition which is good for us consumers

[surreal mantle]

Would it be possible to implement a system where you can add onto an exisiting song but instead of extending you add to it before?
Like setting a timestamp that the AI has to aim to sync up with at the end of the generation?

I've lost track of how many times I've gotten bad generations on the first verse just for the second verse to be an absolute banger which I never get to use and I have to settle for a comparatively mediocre verse where the first verse generated properly or spiral into a million credit clip to get close to that first generations second verse.

[limpid moat]

There is a timestamp to continue from

[surreal mantle]

Yes, continue from.. But that doesnt work when you want to add the first verse ahead of the verse that actually generated proper

[tough prairie]

@feral aspen Need to have an "advanced mode" or musical performer to be more creative brackets reading interpreter [ ()] it describes better each structure, many times when alghorytm check multiple words reads as lyrics or just ignore structures, that will make game more dynamic, creative and useful for musicians or people who has a bit knowledge about music to compose specific music "behavior", for example, if I put a [Violin solo (stacatto, fast-paced arpeggios,etc)] or [guitar solo (acoustic 12 string guitar fingerstyle, E minor, 140 BPM A F#m C#m)] etc

Also initial master instructions like tempo, instruments, language, key, time signature, etc ordered with brackets over "Lyrics"

It will makes more dynamic and less generic sound/ mainstream / top 40 / catchy soundalike

[fervent spade]

Would also like some good Frenchcore training on the new suno Model, I need my brainrot music unless anyone knows how to get 200+ bpm with distorted glitchy kicks

[soft harness]

Would it be possible to have a channel (e.g. #trendy-songs) that Suno automatically submits the links of a song that reached the top 5 or 10 songs of the Trending list so we can look up the song once it is off the page?

Should be a simple web hook with the message being the URL to the audio

[torpid hollow]

I saw a video about Udio, seems interesting. Hopefully its worthwhile for free users

[finite idol]

For the explore page, made by new users might be a nicer category than the weird bottom one there is now

[vague oak]

Explore Beta Feedback

1. Idea great to understand how the sounds look like
2. to spin the weehl ok
3. but… you can Not Control the weel. I founa good sound and then gone
4. sad not combinate different sounds together

[granite current]

I’d really like a search function, where you can search for genres or specific songs on the public page, I wanted to listen to “headlights” but it moved from the homepage and the explore page interface changed

[cedar fulcrum]

Is there a way to keep the vocals /tune and be able to edit the lyrics? Sucks when you get a banger but need to change a line or verse 😦

[paper maple]

Inpainting is a must at some point. I'm spending a ton of points to redo over and over because it decides to rush or even jump around the verse structure.

[safe ingot]

Can you add at least an option to have "extend" look at more than just the previous 60 seconds of the track for context to continue off of? It very clearly forgets anything that was more than 60 seconds back and makes it difficult to work around currently.

[tight beacon]

Please let us select the Image from our "extended" clips I'd rather have a Machine skull (from the first clip Gen) representing my Deathcore Metal than fluffy rainbow clouds from the final clip generated. I beg you.

[regal vapor]

A way to create an account that lets you listen to music, but not make any, that way you don't have to connect Suno to any accounts you care about.

[tight beacon]

Music Generation Suggestion: let users define weather effects in their music like: Thunder, Rain, Wind, etc.

I just had a gen randomly add Thunder in it but I'd like to tell Suno to add it in myself.

[feral aspen]

Rain and wind can sometimes work in the style prompt. But maybe needs some words like "field recording" in style too to work more often.

[safe ingot]

When making any sort of rock or metal music, the vocals always seem to be too loud and make the guitar tracks fade in and out and be hardly even audible. It sounds like massive clipping, even though it isn't a real recording. I don't know if this is fixable or not, but it ruins a lot of songs

[mystic niche]

Train the AI to do better Rap/Hip Hop music. It's always singing and not actually rapping 😂

[frosty lotus]

could be cool to generally give musicians the freedom to expand the base preferences of the song

giving the option to force songs in a certain key, bpm, and support basic transition effects across volume (fade in/out, crescendo, decrescendo) and pitch across the song could unlock a lot of value for technical musicians who can create something more precise much faster

explore should be searchable by genre if possible, but absolutely loving it so far

should def allow songs to be saved that we find on the explore page, this dakar house song im listening to is fire

[leaden marlin]

V4: definitely that background white noise needs to go, ruins pretty much all heavy drops, also we need more control over specific parts of the song, i know you can select continues, but is more about the vocals - where they start, where they repeat etc., also the model currently selects it's own melody - maybe an option to feed the input with preselected synth melody from FL it can incorporate?

[urban hollow]

please insert a subscription via PayPal

[wind abyss]

using of headphones, it still come out the left ear piece, is there a way we can get it equalised to both ears.:) Thankyou

[leaden marlin]

This I can agree with

[worldly pagoda]

Is there will be a possibility to train on a specific singer style? Let’s say I like the style of a specific singer, I want to listen to him singing my lyrics, in his style and tones, but Suno will do the composition and make sure that the singer style/tone/voice is applied.

What I am doing at the moment, is cloning the voice of the singer I want, I make sure my Audio of the singer contains the most tones/chords of his voice (to make sure it covers Suno’s vocal tones/chords), then separating Suno’s vocal from the singer, transferring the cloned voice on top of the separated music based on the vocal provided by Suno.

It would be much better if that could be done directly within Suno. Like optional feature to add a specific singer input then we can upload ONLY VOCAL stem of a singer, the audio file could be 5 ~ 10 minutes long, where it contains the singer voice only in different tones/chords/shapes, to correlate with Suno’s vocal tones. Then it is applied to the music and replaces Suno’s vocal.

I hope I explained it right …

[leaden marlin]

Suno should take advantage of the advantage you have and add some things that Sonauto has such as downloading the songs in stems, uploading a reference song and making songs from any artist

[spiral marsh]

😜
https://suno.com/song/127dfad9-43b8-4617-a249-03bf2e143171

[dapper bay]

Mmmm I don't think it's a suno problem, but a problem from your side, unless you are having a bug, in which case you should report it in the bug channel. I think it's a problem with your headphones or sound settings on your pc. Are you sure your headphones are working properly and don't have a problem like a broken cable or something (if they are wired)? Can you listen to youtube or spotify on both channels? Does this only happen with suno? If they are wired headphones, is the aux connector fully connected to the port and not just half way?

[timid night]

I suggest allowing us to link to a time in a song using the url, such as youtube does. To show a specific part to people without having to tell them to jump to 1:23 would be able to just do: sunsonglink/?s=83

[toxic trench]

I wholeheartedly support the concept of incorporating a feature that allows users to train Suno Ai on a specific singer's style, as suggested by @worldly pagoda This feature would enable users to create unique and personalized compositions that reflect their favorite singer's voice, tone, and style.
Currently, users have to manually clone the voice of their desired singer and apply it to Suno Ai's composition, which is a time-consuming and complex process. By integrating this feature into Suno Ai, users can streamline the process and achieve more accurate and efficient results.
To implement this feature, Suno Ai could provide an optional input for users to upload a specific singer's vocal stem. This audio file could be 5-10 minutes long, containing the singer's voice in different tones, chords, and shapes. Suno Ai could then analyze and correlate the singer's vocal tones with its own vocal tones and apply them to the music, ultimately replacing Suno Ai's vocal.
This feature would provide numerous benefits to Suno Ai users. First, it would enable users to create personalized compositions that reflect their favorite singer's style, enhancing their user experience. Second, it would save users time and effort by streamlining the process of applying a specific singer's voice to their compositions. Finally, it would demonstrate Suno Ai's commitment to innovation and user-centered design, further solidifying its position as a leading AI music composition tool.
In conclusion, I strongly urge the development team to consider implementing this feature into Suno Ai. It would not only enhance the user experience but also showcase Suno Ai's commitment to innovation and user-centered design.

[torpid hollow]

reminds me of this https://www.youtube.com/watch?v=BdOtR2A5agc

[dapper bay]

I want to give my suggestions based on this [https://suno.com/song/28ba86fc-9207-439c-acbb-aa945b4976ec] song.

It would be great to be able to edit the lyrics. There are repeated parts at the end and lyrics that don't really exist in the song.

As probably already said, but to be more specific and be able to ask for a certain BPM, male or female voice, maybe even tonality or mode. And obviously be able to change it or keep it as the next parts or sections are generated.

More consistency, specifically in the voice both in melody and type of voice (male or female). At the beginning it was a male voice and at the end after a long instrumental interlude I couldn't recover the original voice and it was female. I was able to make it generate the chorus similar at least, which was also difficult (several attempts), it was not exactly the same, but as for the rhythm of the chorus melody it was similar enough and it also started after beat 1 which was important as in the original chorus. The melody itself was different but in the end I liked what suno generated and left it that way, after several attempts. It's like suno doesn't remember the original chorus and also that it's a male voice. I don't know if the song is too long or rather, that last chorus or that voice is too far "away" from the last time there was a voice in the song. There is a difference of a little more than two minutes between voices, the instrumental interlude, practically a whole generation. So I thought maybe suno doesn't remember or loses the original info for some reason?

I don't know if this makes sense, but that's the sense I made of it. Although, that instrumental iterlude was generated with a similar but not the same description, so maybe by changing or adding other terms, suno changed or added whatever to his process to generate the next part and forgot or discarded the original chorus? I really have no idea, I just know that something happens. Well, I hope my case is helpful.

[steady scarab]

Jonathan, when can we expect to have something similar to Sora available?

[dapper bay]

No idea if someone has already said it, but to be able to take an already made instrumental song and generate a voice with automatic or custom lyrics and be able to define the style or whatever with the song description.

[median tinsel]

A suggestion that should be taken into consideration: Single voice option, without chorus and choir

[robust jewel]

The new Explore page is a nice idea, but the dark pink/red/purple text on a purple/blue background isn't exactly the easiest thing to read.

[gleaming wyvern]

I will make a new membership. Does anyone know a promo code?

[worldly pagoda]

At least not the style, but the original cloned voice characteristics and tone. because now 100s generated somgs, are similar in (Suno’s default) singers voices. We need more customized voices, we can help Suno in creating new voices if everyone started to train different voices, different styles, different tones and languages. Look at the huge potential in here.

[random adder]

I was just on the Suno webpage and Wish their was a way to comment my thoughts on each song (notes to help with improvements)

[wise pine]

"insert metal yell" tag?

Also i think suno needs more training on Symphonic Power Metal

[silk rose]

I need the ability to end a song at a certain time, not just keep it going. Some songs feel like the should end at a certain point but they drag it on for 30 seconds longer

[final cairn]

This has probably been suggested but I when going past 3 or 4 samples the quality severely degrades and the overall sound & mood begins to drift

[regal bone]

Please put variety in the voices and music again and remove the graininess or layered computer sound of the voices.

[hollow locust]

Best work around for now is 'get full song', then make another part 2, repeat.

[remote leaf]

The model that generates the music does not respect line endings sometimes. For example, the lyrics
Roses are red
Violets are blue
Sugar is sweet
And so are you

The music generated might sound like
Rooooseeeees.....
Are red violets....
Are bluuuue sugar is sweet....
And so are you

[formal bear]

yeah I hear you

will probably be on demand as well, but more options

[swift parcel]

The V3 alpha model excelled at expressing underground music and inspiring musicians, but the V3 model produced more predictable and mundane melodies, lacking diversity. In particular, it fails to properly capture the essence of underground band music. If this trend continues, there are concerns that talented musicians will leave the platform.

[solid dirge]

Exactly what I came to suggest.

[jaunty moat]

Second on the search function. Search is a pretty standard and much needed thing for something that has a "library."

[hollow locust]

For 'Part 2' to also be 02:00 mintues long.