community-help

[hard cobalt]

no

[timber shadow]

When you tried in an incognito/private window, were all browser extensions disabled?

[hard cobalt]

yes

[timber shadow]

Since you don't have a VPN installed, do you have any other program or configuration that could be filtering your traffic? Antivirus, firewall, or even DNS settings on your device or internet network?

[hard cobalt]

nope, i have windows defender but the settings have not been adjusted or something, and I am using a lan connection, no wifi connection. No program or configuration that could filter my traffic then

[timber shadow]

Do you have a WiFi card you can connect with, or LAN only? Just wondering if you can test on another network like a smartphone hotspot. If you can, be sure to try again in a private window, no extensions.

[glad hull]

ah why does chat gpt only work on phone and not browser, I tried restarting and all browsers T_T

[brave furnace]

What is up with the GPT-4 "network errors"? I am getting them constantly on GPT plus, never experienced this issue for months prior (think it is because of larger coding questions and longer responses), but it simply cannot finish a reply without having a network error. I have 1Gb/s speeds on and am connected via ethernet, have tried on numerous desktops/laptops, on numerous browsers (also tried clearing cookies & cache) and have been on two diff networks that are both extremely fast yet nothing changes this network error

[signal fern]

/video

[timber shadow]

If it's working on one device and not another, it could be a VPN, browser extension, browser setting, antivirus/firewall program, DNS setting, or some other configuration thing causing the issue.

[patent crag]

same problem here

[west sail]

Has anyone noticed if GPT 4 is operating in a Suboptimal Mode? The whole system has been demonstrating uncharacteristically deficient performance with even basic prompts, evidencing a concerning regression in abilities. It appears unable to execute straightforward tasks one would expect proficiency in, like summarizing a simple PDF. I would greatly appreciate any insights into the factors that may be impairing GPT 4's skills and inhibiting it from reached its previously exhibited potential. Please advise if you have also observed this pronounced decline in GPT 4's competence at fundamental capabilities it has historically shown aptitude for. It remains unclear why its intellect appears so diminished presently compared to its prior exhibited mastery when responding to basic language prompts.

[tawdry minnow]

hi I saw the "Our systems have detected unusual activity from your system. Please try again later." message when I talk with chatGPT-4. And I cant fix it.I just came from Turkey to study in New York, I already have another account in Turkey but I bought a new membership here. because I saw the same problem in that account and I thought it was not used here. But even though I bought it new, I'm still having the same problem, please help.

[reef zealot]

keeps giving error when using gpt4 but not gpt3.5

[loud dock]

Hi i renewed my subscription to chatbot 4 but it didn't subscribe I'm still on free version

[fervent crown]

dall-e on firefox opens the main page to write a prompt, but after typing in a prompt it goes on a white screen and never shows the results
reloading the page shows the results for 0-2 seconds before disappearing again.

disabling adblocker does nothing, disabling (stylus) styles does nothing, and none of my other extensions interact with the webpage at all

this issue is persistent across all of the browsers i have used:

• firefox
• microsoft edge
• google chrome (as of 2 months ago)
• safari (through my iphone)

chatgpt works fine, but dall-e doesn't for some reason

[unreal hemlock]

<:book_icon:1171408210398289941> `` Rule 2 `` Keep it G-rated.

Do not share or display content that is not G-rated, including profane or vulgar language, via any means. NSFW or violent/disturbing content is strictly prohibited.

[untold zodiac]

Any reason why I'd constantly be getting this error?

My apologies, but it seems there's an issue with accessing the file you've uploaded. I'm unable to use the myfiles_browser tool to open it and provide a title. Could you please try re-uploading the file, or provide more context or content from the file so that I can assist you in creating a suitable title?

this is my current flow, creation is done in parallel in separate threads
weird thing is, creating flashcards may work, but creating mindmap or title wont work.

[thorny cloak]

Once upon a time, in a dense forest, there lived a small animal called a little fox. The little fox has a pair of dexterous paws and a brave heart. However, its biggest dream is to become an excellent fox dancer.

[whole dock]

Hi guys,
Is there any way I can speed up the API responses from GPT-4? The input is usually about 1000 tokens long but the response takes upwards of 7-8 seconds to come.

[patent abyss]

Hello, I can't login to my chatgpt Plus anymore, it says "You do not have an account because it has been deleted or deactivated. If you believe this was an error, please contact us through our help center at help.openai.com." Please could someone check? thanks

[past geyser]

You can try the turbo model if you're on API - https://platform.openai.com/docs/models/overview

[past geyser]

The message seems pretty unambiguous unfortunately. Nobody here has access to your account information.

[austere prism]

I'm receiving the "conversation key not found" error again as of this morning with chat history turned off anytime I try to send a second message in the thread

[full surge]

the scrollbar of the y axis of the prompt not work anymore what is the issue ?

[full surge]

really it is so boring i need a help, when i write my request in the prompt the prompt go bigger and bigger and i can't see the other question and answer have been received before. last day it worked well and the scrolbar was active but now the scrol bar not work and there is any scroll bar in the y axis ,

[timber shadow]

Just tested on the web version, I also don't see a scroll bar anymore, messages just seem to expand indefinitely.

[full surge]

is there any solution ?

[timber shadow]

It might be intended behavior, so might not be something that comes with a solution. I see how covering the conversation isn't ideal always, but I can also see how a small text entry box could be suboptimal as well. You might consider drafting very long messages in a separate text editor, then copy/pasting it over so you can easily reference the conversation when writing a longer response.

[cold folio]

it should be auto

[full surge]

unfortunately not auto and when i changed it to auto it reset to default mode automatically , is there any solution ?

[cold folio]

What issues do you get at the moment ?

well i just got one now 😄 if the prompt is too big, the scroll bar does not follow the end of the prompt or current state

[full surge]

i want it to be like before when the data and question in the prompt go bigger and bigger the last questions and answers in the same prompt go disappear and you cant also scroll . was it clear?

[cold folio]

yeahi also got an issue right here, if you use backspace, the view goes to thet top

it wasn't like this 2 days ago

[full surge]

the first lesson of the web development is the text area should not go to infinity but here in the Chatgpt this simple thing has been ignored

[cold folio]

i'm not a fan of tailwind as well. when its time to debug, it takes so long that its just better to have their own custom css

[safe notch]

guys, for some reason my ChatGPT 4 is not respondidng at any of my questions. anyone knows how to fix this?

[timber shadow]

Hello, are your messages just disappearing when you send them? Can you share a screenshot?

[weak ruin]

is the dalle 3 api down?

[safe notch]

yes, just just disappearing when I send them

[tame depot]

Hello! I have trouble using whisper on my computer, can anyone help me? Am I at the right place to ask questions?

[safe notch]

@timber shadow yes, just just disappearing when I send them

[weak ruin]

https://status.openai.com/

its down

[timber shadow]

What browser(s) is it happening in?

[safe notch]

@timber shadow is Chrome

[timber shadow]

That looks like the Vision function, shouldn't be affecting DALL·E.

[weak ruin]

mhh

[timber shadow]

Can you try in Firefox as a test?

[safe notch]

oops, not really as firefox is not allowed

[weak ruin]

not allowed?

[timber shadow]

Can you try in a private window of Chrome then, with all extensions disabled?

[safe notch]

will do that

nope, not working

not allowed as per company policy

[halcyon ferry]

Its jus t blank

[timber shadow]

It could be something interfering on your company's network then. You might try reaching out to your IT for support. As a test, you could try logging into the web version on your smartphone, connected to a cellular network. If it works there, it's likely the company network/extensions interfering.

[timber shadow]

Can you try in a private window with all extensions disabled?

[halcyon ferry]

I found one fix, when I use VPN it is working fine, but without VPN its blank

[weak ruin]

why is firefox not allowed

[weak ruin]

where you located

[halcyon ferry]

same, its blank in private window as well without VPN.

[halcyon ferry]

India, I am using ChatGPT from more than a year now. Also have the teams plan

[weak ruin]

maybe it's the issue that they blocked it

[tame depot]

So I have a Mac laptop from 2017 (MacOS Monterey 12.7.1). I have installed Python 3, Pytorch, ffmpeg, pip, and whisper via pip. Every one of these installation processes seem to have worked well and have been fully complted. But when I lauch "whisper [path/to/audio/file.mp3]" it says "zsh: permission denied: whisper". I have tried to run my terminal with root permission, but even then it says "-sh: whisper: command not found". Can anyone help please?

[timber shadow]

Maybe try in #dev-chat if you don't hear back here! Some helpful devs in there.

[safe notch]

thanks guys, much appreciated!

[halcyon ferry]

I can't understand this, as soon as I messaged it here it started working as normal. But from the past 5 hours I am trying and it was not working.

[weak ruin]

mhh idk

[halcyon ferry]

Anyways thanks for your input guys!

[rapid flower]

When did this become a thing? I never get complete answers anymore. It is super vague, and now it will not even produce a simple picture. smh...

[timber shadow]

Star Wars is trademarked IP, which DALL·E can't produce images of.

[rapid flower]

Hmm... I guess I have never specifically tried using IP content before. Thanks for the answer.

[weak ruin]

someone can check the #dev-chat

[sullen raft]

guys I ran into another issue

I was using gpt 4 as normal

and then got hit with an error unusual activity detected

and then I am no longer able to send messages

weird enough is that there is not any unusual activity so idk whats going on

now it works

[halcyon ferry]

Try deleting the site cookies and cached data and then try

[hidden elm]

👍

[mystic vale]

An unusual question maybe: Is there an official ChatGPT merchandise store (that delivers to the UK)? I would like a silver-metallic lapel pin for my suit sporting the ChatGPT logo (not a plastic button-style one).

[versed dagger]

Hey guys, I need your ideas!

I want to develop an AI shopping assistant that can recommend products. So he needs all products with all relevant information from shops, this information is in the format seen in the image. I'm using the Assistants API and this txt file as the knowledgebase.

The Problem: The assistant takes long to respond and the information is often not completely correct.
So my Question: Which solutions do you know that could fix this, apart from Fine-Tuning?

Really appreciate your help! 😄

[inner brook]

Is ChatGPT down?

[valid shadow]

is GPT down got team account, getting We ran into an issue while signing you in, please take a break and try again soon.

[rigid musk]

status.openai.com

[rugged wing]

Yes, the https://status.openai.com site just flagged an issue with ChatGPT

it should be fixed now

[bronze charm]

Today's incidents are still marked unresolved; however ChatGPT is working fine for me.

[rugged wing]

is what I'm seeing

[bronze charm]

Ah, was looking here just a bit below

[rugged wing]

from chat.openai.com? @livid acorn

I was not aware you could buy teams monthly

the only times I've interacted with users they've told me is was $600 billed annually

[rugged wing]

right, so it could be 4x30$ + taxes .. I wonder if one of the other users has invited someone else

I've not had chance to use teams myself yet, but I seem to recall being toild that other users could invite new members

if you click on your name in the left hand panel you should see a popup that says "My Plan"

bottom left is an icon and your email

then a manage my subscription option at the bottom of the curent plan

right, i get you, I belive you can request that your old account get rolled into your teams account if you wish

[livid acorn]

any tips on how to get them to reply to me? I've sent a couple help requests, and they said 3 days, but it's been 6

[rugged wing]

details here https://help.openai.com/en/articles/8801890-can-i-migrate-or-merge-my-chatgpt-free-or-plus-workspace-over-to-my-chatgpt-team-workspace

[rugged wing]

unfortunatly they are super busy at the moment, it can take several days to a week to get a response at times, hopefully that link can guide you to merge your old account with your teams account, you can always try connecting again to https://help.openai.com and sending another message

[livid acorn]

Okay, I really appreciate this. They have a strong discord community, so this helps those who fall through the cracks with CS. Thank you again!

[rugged wing]

[glad hull]

still unable to access gpt premium myself : (

[arctic holly]

What’s the point of being a paid member if your giving new features to free members to test and now paying members?

[rigid musk]

browsing? gpt 4? dalle 3 ? vision? and more

free users doesnt have those option

it isnt the biggest feature

[arctic holly]

Memory is a huge feature

[golden socket]

what is this?

[twin narwhal]

Are plugins not working for anyone right now

[ocean oak]

No,, it says no plugins installed, then when I search for plugin it says uninstall, but it won't let me uninstall

[patent kernel]

For those of you that are wondering if you sign up for a teams account you can merge your personal account into a teams account.

Also does anyone know how we can check to see if we have access to new beta features with team accounts. We do not have the beta section under settings. Like when the @ mentions launched there was no place to turn it on. It just started working.

[vivid mortar]

Helpful information, thanks for sharing!

[brave furnace]

I believe it automatically "shows more" on analysis and other stuff. Normally you have to hit the down arrow but this will automatically show it as it is in progress and afterwards.

[fallow snow]

Any tips for dealing with hallucinations of unable to fulfill

[fallow snow]

Like my chat is completely stuck

[bleak elk]

I want to use the Chat GPT team plan with a friend, but can we designate both of us as owners? We’d like to be able to make detailed settings individually.

[livid nexus]

would any of you know, how I could change my profile mail address as I am leaving the company I used the address of?

[limpid musk]

same for me, I need to change my profile name to make it better for GPTs. But not able to do it.

It's like inherited from my payment card.

[rugged wing]

There is currently no way to do that unfortunately.

[rugged wing]

you can register a domain to use as the name for your GPT

e.g.

[livid nexus]

ok. so I would need to quit this subscription and assign for a new one then?

[rugged wing]

yes

you can have your phone number assosiated with up to 3 accounts, so assuming this is your only account with that number you should be fine

[vital pecan]

does anybody here know about any opensource model that can help me transcript letters and characters (speech to text). I am willing to finetune it. Also, will finetuning whisper will help in this regard?

[rugged wing]

Sure, you can fine tune whisper and it's open source

I've DM'd you a medium post regarding it

[vital pecan]

Alright Thanks

[meager rune]

can we not invite ChatGpt to our own server?

[rugged wing]

#gpt-discord-bot was the last thing I saw that worked, not sure if it's been kept up to date, worth checking out

[meager rune]

It says no access to me

[rugged wing]

I've DM'd you the github link that channel contains

[meager rune]

Thanks i just saw it

[elder cradle]

Software gore...

Seems like it tries to call functions to change the GPT, but they fail

[lunar basin]

I want gpt to tell its name Agex

[halcyon ferry]

you may train your own GPT to achieve that.

[lunar basin]

Sorry I mean i want it to remember my name

[halcyon ferry]

Oh, then you may add that in the custom instructions.

[lunar basin]

how

example please

[halcyon ferry]

Here, click on the profile section, and then click on custom instructions. And then add your details about you in the section that opens up

[lunar basin]

ok

thanks

[cold folio]

Hello. Does it happen with high frequency or just happened once ?

[elder cradle]

Happened now, but I haven't tried making a GPT in a while, I'll do some more testing

[red flower]

many 404 errors

[Error] Unhandled Promise Rejection: FatalServerError: Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
(función anónima) (_app-226fb5ab17dbc028.js:60:119802)

[lunar basin]

but wai

wait I am talking about api

[blissful latch]

I think your API ist dead.

[amber steppe]

whole thing is broken RN??

[full surge]

thanks OpenAI for taking money and returning error without any explanation to users

[rugged wing]

thanks OpenAI for not having a working login system

[sleek pulsar]

I have the same error. It'll get fixed

[opal kayak]

me too

[limpid musk]

Same happens for me

[somber karma]

Getting an "error in moderation". Anyone else?

[median charm]

yeah

[carmine wigeon]

Same here

[hoary zephyr]

yep

[median charm]

seems like the backend is broken

[somber karma]

Incredible!

[median charm]

oh god, I forgotten how to write code!

[elfin dirge]

Is the whole ChatGPT down?

[placid fern]

yeah it's down for me here too weird how the status page shows everything as being online but it seems like it's down for a lot of people.

[cosmic agate]

oh man...my job is gonna take longer to do today if this ain't up xD

[teal pier]

has chatgpt again some problems bc it isnt loading for me?

[elfin dirge]

That is funny, I have plus, so it could at least work

[elfin dirge]

Good it isn´t an issue on my side

[median charm]

guys help! be my AI for me! I've got a website that ive dev'd locally, and want to deploy it "properly" with docker CI/CD, etc. give me useful advice that i need to follow up with credible resources T_T

[rugged wing]

hot take: prompts by plus users should be on seperate hardware

[median charm]

oh well. looks like I'll have to use YT, blog posts, etc.

[placid fern]

Agreed

[median charm]

yeah, I'm on + as well.

[placid fern]

agreed

so am i

All I want is a code review haha

[median charm]

...my guess is that's not the issue. if they haven't distributed their backend, it would be trash-teir infrastructure.

[somber karma]

I have two accounts. Neither is working

[cosmic agate]

it works then gets cut halfway through lol

[elfin dirge]

I tried both models, and not a single one worked

[rugged wing]

i dont even have plus but any good service should at least give some soft of better service or seperation from the rest of the users

[elder cradle]

Now this:

[placid fern]

The error in moderation scared me i was trying to figure out what the heck i did lol

[median charm]

issue is probably a part of the system makes it a bit of a house of cards, and an intern has been given priviledges that they shouldn't have been given.

[rugged wing]

hmmmmm. it seems chatgpt pulled another chatgpt

[placid fern]

or someone tripped on the server power cable. haha

[timid igloo]

what does that mean?

[elfin dirge]

I have this EXACT SAME issue

[elfin dirge]

same

[rugged wing]

"how github went down for 48 hours after 43 seconds of downtime"

[severe owl]

Been having issues with chat gpt loading/ generating responses for the past two days. Doenst work at all

[timid igloo]

Yea, I have met this issue many times a day...

[elfin dirge]

Also it is very sad that even when I pay 25 dollars a month, the customer support is just... meh

[queen forge]

Does anyone know how long such errors last?

[elfin dirge]

There was a Dall e error some time ago, it lasted for I think not even a hour

[elfin dirge]

I started having issues today

[placid fern]

well considering it doesn't show on the server status page who knows if they're even aware yet

[celest edge]

Boo

[elfin dirge]

Apparently they know about it and are investigating it

[cosmic agate]

i hope they fix it today.

[elfin dirge]

Me too, if not, they should be ashamed of themselves

I pay for plus but yet am unable to use it due to THEIR issues

Fixing it is the least they can do

Ok, expect a ping in 56423 business years

[hushed phoenix]

dang, I literally joined the discord to see if there was any solutions for that (error in moderation) haha

[elfin dirge]

They just MUST fix it

[cosmic agate]

same here lol

[elfin dirge]

If it is down for longer than 5 hours then I ask for a refuns and I don´t care if I have or don´t have the right to get one

This is me rn

I wonder why do they not give refunds for errors, downages and other issues

Oh wait, I know!

They would go bankrupt in one day if they did that

Yes, I am pretty sure the leadership is laying in a bathtub full of money not caring about the users at all

What do you mean by +1

[hardy orbit]

Lukas you're complaining about $20

A month

[queen forge]

that's a lot for some

almost a week of food

[meager ridge]

^

[hardy orbit]

Plus is also optional lol. If it's too much then just use 3.5

[elfin dirge]

25*

[elfin dirge]

In my country hit by inflation and everything is expensive - every money is good

[vagrant plover]

Tip: you can subscribe to updates on the status page, letting you know when things are back up and running right away

[queen forge]

of course it's optional but it's still a bummer, if you payed for it and it's not working like it's supposed to

[stoic bobcat]

hello im encountering a connection error and also a moderation error

[ashen tendon]

As we are !

[queen forge]

we do too

[stoic bobcat]

okey i tought i was alone

[elfin dirge]

Czech republic

[stoic bobcat]

morocco

[elfin dirge]

Oh hey brother!

Polish and Czech are brothers

[queen forge]

[elfin dirge]

really?

[stoic bobcat]

really ?

[elfin dirge]

Finally

[queen forge]

it's generating results for me

[elfin dirge]

Not for me

[stoic bobcat]

still error in moderation for me

[elfin dirge]

It is not on the status page as well

[elfin dirge]

same

Because "fair" rules

[burnt scarab]

I keep getting errors with my personal plus plan and business workspace account. 😦

[elfin dirge]

everyone does

[queen forge]

still doesn't work, had luck with one

premium

[elfin dirge]

I try both and not a single one works

[queen forge]

hm, I'll try but I'm using it for a research paper and already got a lot of information in my chats

[plucky plume]

same "Error in mModeration" error on free and paid version

thanks guess that would explain the slow death i experience 😂

[queen forge]

it's been an hour

[sleek brook]

hmm yes gpt-4

[queen forge]

funny, default doesn't work

[sleek brook]

It doesn't, just gives the same message

[queen forge]

they should include more error messages, for fun and entertainment

[elfin dirge]

Guys I think it works

[elfin dirge]

yes

[flint karma]

hello i wnat to fix this
if i click " use default model "
he send it agin
help plase

[queen forge]

the bot is down, currently doesn't work properly

[flint karma]

ok ty

[sweet zinc]

I am not able to upload files to the bot

[deep grail]

same issue

[formal glacier]

Does anyone know how to contact someone at OpenAi?

[velvet heron]

Where is that nifty link to check if android version is down? For some reason it's down for me in past 2-3 hours.

[rigid musk]

Help.openai.com

[formal glacier]

Tried that, but the chatbot just keeps giving random options, or "submit a request". Is there not just a way to actually live chat with a real person?

[rigid musk]

You have to fill a request and you will get back by email

[raw hazel]

I'm getting a blank screen after logging in on safari desktop

[rugged wing]

that does not sound like part of the same problem

re-boot/reload of the browser would be worth a try

[median charm]

seems to be working again. anyone else running into problems still?

[next orbit]

Still not running for me right now.

I've tried multiple browsers and Chat GPT won't show up.

There have been large spikes of outtages on down detector since 6 am PST.

[tropic lily]

Has anyone worked on .dwg files and OpenAI? I am creating a custom GPT and need assistance. Please DM me.

It involves analyzing the interior structure of a commercial building for specific industry calculations

[rugged wing]

I'm having issues uploading images to my GPTs. When I do it locks up Chrome and I get a popup from Chrome asking me if I want to close the unresponsive page. Is this happening to everyone?

[velvet knot]

Trying to get access to finetune GPT-4. Been on the waitlist for about a week. Anyone know what the wait looks like?

[ancient pewter]

why did chatgpt output something unrelated to my prompt

[bronze charm]

the prompt of all time

[blissful latch]

ya its still not working. My api still tells me that GPT is not able to open PDF and docx files.

'utf-8' codec can't decode byte 0x90 in position 14: invalid start byte

[rustic marlin]

some people reported the AI doing that, but I find hard to believe that you didn't primed the AI to reply in such way

could it be related to the newly released memory feature?

maybe it is primed with this sort of joke on its memory, you can check it on Settings > Personalization

[ancient pewter]

i did not prime it

[rustic marlin]

what about memory?

[ancient pewter]

my gpt is free and idk if i have access to memory

also the automod took my previous message out of context and muted me for a hour it is a false mute i think bc idk that there is anything wrong with it

[verbal igloo]

Quick question - Any idea how is this metric calculated? unique conversations/ cumulative sum? Not able to get good insights from OpenAI Website. Thanks

[timber shadow]

Hello! It's number of chats that have been started with the GPT, your own included.

[rustic marlin]

free users also get it, check if it rolled for you, click the settings menu > Personalization

[verbal igloo]

Thanks! Then this is then very much a gameable metric (provided folks have a premium account)

[ancient pewter]

i did not get it

[rustic marlin]

then it is likely just a random weirdness

some people reported those weirdness

[ancient pewter]

i get those sometimes

[mossy brook]

Hello, I'm new to OpenAI, how can I purchase a Gpt 4 Key?

[rocky rover]

I provided you the answer in #1207514554188308533 message

[mossy brook]

Thanks!

[opal kayak]

Something went wrong. If this issue persists please contact us through our help center at help.openai.com.

[tidal stump]

Hi everyone,

I am trying to build a custom Assistant for retrieval from structured data (xlsx, csv).
It seems that Knowledge Retrieval doesn't support xlsx, csv files.
What's the best way to build an agent that can answer questions from tables?

[proper dirge]

Hi.
function_call is not calling in most times specially in nexjs application. any one having this issue ?

[valid panther]

I used an online guide to make a raspberry pi GPT voice assistant using a free subscription account with a few credits purchased but it doesn't have live internet access. Would upgrading to a paid subscription enable me to change that?

[rigid musk]

They are no subscription for the api

[valid panther]

Ahh ok, so is live internet access not possible with the API for a clueless muppet like me?

[rigid musk]

It is pay per usage for the api. It money isn't an issue for you. It will work for sure

[valid panther]

The voice assistant works, it just doesn't have live internet access and I'm trying to find out if I can make it so it does.

[astral vessel]

Hi guys, I am facing this issue, which custom-GPT wont connect with my server.
Which I use full link directly in browser it works.
Error from custom-GPT config and complete server URL given

[bright hare]

I'm wondering, if I have the paid ChatGPT, can I upload a few pages (photos) of my textbook and get it to generate questions for me (to check my understanding) based on those pages?

[pulsar tapir]

Hi, anyone knows why DALL-E stopped generating images with GPT-3.5? I'm a Plus user, but GPT-4 has a usage cap whereas GPT-3.5 hasn't and I've been able to generate images there. The possibility dissappeared about a week ago. Can someone confirm that?

[rustic marlin]

im pretty sure it was a bug

[dusk shuttle]

GPT-4V can currently only recognize English well in images. If your book is in English, then there is a very high probability that it can do it! It can also understand more complex graphics in images to a certain extent, but please ensure that the images you submit are clear enough, and forgive the web version of GPT-4V for only being able to receive 720P resolution images sent by the server! However, for the specific settings of the question, you need to communicate further with it!

[bright hare]

thank you very much!

[dusk shuttle]

😄

[spice gate]

guys, i need help, I cannot login to my gpt account, it says my email isnt supported, I have been using the same account for over half a year now, and its login with apple

anyone know what is going on with gpt today? the same account can be logged into my ios devices but not the web ones'

[pulsar tapir]

What exactly you think was a bug? The possibility of GPT-3.5 to generate images with DALL-E? Or what I have now

[rustic marlin]

I think 3.5 wasn't supposed to run such function calls

recently a user commented that they where able to do so and a few days ago they commented that they where not able anymore, so, I can only assume that this was unintentional

I think Plus memebrs can still have 3.5 do function calls if it is a chat that originally started with 4.0

(Im not sure about that tho, someone please rectfy me)

[timber shadow]

Hello, just to clarify: when you are able to log in, are you clicking "Continue with Apple" or are you typing in your email address and password directly?

Edit: nvm looks like you got an answer here #openai-chatter message

[spice gate]

yes, i am using continue with apple, i am not able to login, just on chrome, i tried safari and it worked

[proper ether]

I'm a GPT PRO subscriber and I tried to use it today, but only the message below came up and it's not progressing
Does anyone have the same symptoms as me?

[timber shadow]

Hello, do you have any unofficial browser extensions installed that might be affecting ChatGPT? The logo next to "ChatGPT" in your screenshot looks a little off center for a GPT-4 chat, here's a screenshot I just grabbed from mine for comparison.

[proper ether]

Nope\

I think that's the case when you zoom in on the page

[timber shadow]

I see! Is this happening with both 3.5 and 4?

[proper ether]

Unfortunately, yes

[timber shadow]

Can you test in a private window with all extensions disabled?

[proper ether]

Same

I don't understand..... 😭

[timber shadow]

Do you have chat history disabled?

[proper ether]

No, I thought it was because the records were piled up, so I cleaned the chat history.

[timber shadow]

Are you currently trying in Chrome? If so, can you test in a non-Chromium browser like Firefox as another test?

[proper ether]

wait

It's the same on Edge, but I've never used Firefox

[timber shadow]

Edge is also a Chromium browser, so if it's got anything to do with that, Chrome and Edge might show the same error and Firefox might not, because it's not a Chromium browser.

Also, do you have a VPN installed?

[proper ether]

nope

[timber shadow]

Okay, are you able to test in Firefox?

[proper ether]

...;(

[timber shadow]

So strange! Do you have another device and network you can test with, like a smartphone on a cellular network?

[proper ether]

The safari is the same

[timber shadow]

Can you test on cellular instead of WiFi? Maybe in a private tab too!

[proper ether]

That's weird

[timber shadow]

Strange! Thanks for trying so much troubleshooting. As far as I know, GPT-4 should be telling you you've reached a usage cap, not a 24 hour cap. There are some posts on community.openai.com I've seen reporting a similar error in months past. Perhaps this is just a temporary bug.

You might consider posting in #1070006915414900886 -- OpenAI staff are periodically present in that channel. Be sure to follow the bug report format in the pinned post. You can also reach out to OpenAI support at help.openai.com. Open the chat widget in the bottom-right, click through the most relevant options, then you'll eventually find a path that lets you type your own message. This will submit a ticket that OpenAI support can review and get back to you on. It may take some time. You'll get an email when they reply to you!

[ripe sun]

hi does anyone have chatgpt + and can check a text for me?

[proper ether]

Thank you, Solbus man

[versed swift]

how can i get rid of "network error" ? i keep getting this

[quasi willow]

yep me too. android app has been completely unusable the last few day because of this

[timber shadow]

Hi again, do you have ChatGPT Plus, or did you have alpha access to Plugins as a Plugins developer?

[nova oracle]

i unscubcribed a long time ago, i got acces as a plugins developer

[timber shadow]

Gotcha, I've seen other plugins devs lose access like this before -- I think it might be related to the fact that plugins are being deprecated as they're replaced by custom GPTs. Perhaps you got an email about this? Or maybe you can reach out about it as a dev.

[nova oracle]

i waill check my emails hold on

[nova oracle]

can it be fixed

[timber shadow]

I don't think it's a bug, I think since Plugins are being replaced by GPTs, Plugins development and general Plugins access will likewise be deprecated. I think Plugins updates have been over for awhile now, for example.

[nova oracle]

so i cant get it back

[timber shadow]

I can still access the Plugins model on ChatGPT Plus, so you should still be able to use it with a Plus subscription. If you have any dev resource contacts related to developing plugins, you might try reaching out there too.

[frank needle]

hi sorry, may i have some help to solve the problems of the chatgpt of my network? its been a headache

[timber shadow]

Hello! What kind of network problem are you having?

[frank needle]

You see, I had been using it normally to study a few months ago, but since yesterday it hasn't given me a response, on the left side I don't get old chats and I always get "something wrong", in the browser console it shows 403 forbiden" and " ....its not a valid JSON"

The thing is that it doesn't work only with my internet, because with other wifi it works correctly, another thing that confused me is that apparently it's not the wifi because it doesn't work either in the chrome app on the phone or in any of my browsers on pc , but in the android application yes, it is quite strange

[timber shadow]

Thanks for all the details! On your WiFi network and PC, what browsers have you tested it in so far?

[frank needle]

microsoft edge in pc and google chrome in pc and phone

[timber shadow]

As a test, can you try in Firefox on your PC?

[frank needle]

any idea?, sorry I know it's a complicated problem.

[frank needle]

i will install it now and tell you

[frank needle]

yes, doesnt work too xD

[timber shadow]

Dang! Thanks for trying. Have you already tried in Chrome on your phone with WiFi off, on your cellular network instead?

[frank needle]

yes!! I used it with my mobile data, which led me to think it was a wifi problem, but since it works in the android app I'm not so sure anymore

[timber shadow]

So just to clarify, it works in Chrome on your phone if you're connected via mobile data? Just not on WiFi, unless you're using the app?

[frank needle]

yes, in the app works in both situations, but in Chrome Phone only with mobile data, if i use wifi on mobile chrome it gives me the same message from all the pc browsers

[timber shadow]

Okay! Thanks again for all the detail, super helpful. Here's what I think is happening. It's likely that the app connects to the services differently than the web version does, which could explain why the app seems to be more "resilient" than the web version when it works in more situations.

In other words, I do think there's something about your WiFi configuration that's throwing it off somehow, even though the app works on WiFi, because of how it works when you disconnect from WiFi. A common culprit in cases like these is the DNS setting of your network. Do you know if you have a custom DNS configuration on your router by chance?

[frank needle]

I was recently following a tutorial and it made me do ipconfig /release /renew etc in the Windows console
Outside of that, don't touch the DNS, would I have to put a specific one? I think this is not the first time that DNS has given me problems.
I can access the internet modem from the IP, I could enter and touch something about DNS if necessary but I honestly don't understand much about the subject, I am familiar with how to do it but I don't know what to do

[timber shadow]

I think the Windows console stuff you mentioned would just be for the Windows device and wouldn't affect your network in general, so no worries there. Instead of diving into investigating the network's DNS settings right away, have you already tried just a simple off and on again of the network? This could help clear up an issue and would be pretty easy to try. If you're able to try this, maybe try signing in fresh in a private window of a browser after WiFi comes back online to make sure old cache stuff won't interfere with the new attempt.

[frank needle]

Yes, I tried restarting the modem quite a few times, since it is usually the source of the problems butttt, nothing, I also tried to do it in an incognito tab and everything remains the same, sorry that the topic is quite heavy but I don't understand either, what it worked bad

[timber shadow]

Dang, okay. Another quick check before checking network DNS settings: do you happen to have a VPN installed on the PC and phone?

[frank needle]

no, i never used a vpn

[timber shadow]

Okay. If you feel like it then, you might check your network's DNS settings. It should be labeled "DNS" no matter what service provider you use, it'll just be a matter of finding it. On mine, for example, DNS settings can be found in Network settings, then behind Advanced Settings, but yours may be in a different spot.

[frank needle]

i got it, should i use a specific DNS or anything is fine? I can enter the code by default as well but that didn't change anything to the problem, I had to enter 1111 1001, any other options?

[frank needle]

or some option that i have to mark?

[timber shadow]

If you've never adjusted the DNS settings, it's likely that it says "Default" or maybe is just blank, which would probably indicate default settings. There may be lines for "Primary" and "Secondary." It's possible that your ISP's default DNS settings are causing some conflict, but this is speculation on my part. If you'd like to test it, though, you might try using Cloudflare's DNS or Google's DNS, they're both popular and some have reported an improvement in connection issues using one or the other of these. You can find more by searching online for what the actual DNS addresses are -- they look kind of like IP addresses and you just type em in and they do their work for you from there!

[frank needle]

I tried it and nope, nothing, I just give up xd, it's very stressful and I've been dealing with this for days, I don't know how to solve it,
It's a shame because it helped me a lot to study, I'm going to try to get used to the phone but it's faster to have it at hand, anyway, thank you very much for being so patient

[restive magnet]

what happen to history?

[bronze charm]

Press the sidebar button: middle left of the screen

[restive magnet]

You mean next to the word ChatGPT?

[bronze charm]

Nope, it's a bit hidden and looks like ">"

Bit below that

In the middle of the left edge

[restive magnet]

Oh I see it damn, very hidden. Thanks!

[bronze charm]

🤘

[timber shadow]

Shoot, I'm sorry, thanks for trying so much. To do some more testing, you can try this:

• Open the dev tools of your browser, usually by right-clicking and clicking "Inspect"
• Navigate to the "Network" tab
• Produce the error on ChatGPT -- you'll see some errors appear in Network
• Make a post in #1070006915414900886 sharing this -- OpenAI employees are periodically present there, just be sure to follow the bug report format in the pinned post!
• Alternately, reach out to OpenAI support at help.openai.com. Open the chat widget in the bottom-right, click through the most relevant options, then you'll eventually find a path that lets you type your own message. This will submit a ticket that OpenAI support can review and get back to you on. It may take some time. You'll get an email when they reply to you!
  Good luck! Hope you're able to resolve this soon. In the meantime, since it seems to work on your cellular network, maybe start a hotspot on your phone and connect your PC to that network -- that should let you use it from your PC! Please let me know if you find a fix! 🤞

[vernal badger]

When an error occurs during a chat and the screen is refreshed, past chat logs may be interrupted or become invisible. (Just to clarify, they haven't been removed by the system through censorship.) In severe cases, the conversation logs may become jumbled and practically unusable.
This phenomenon has been happening frequently in my environment lately, and it often coincides with issues such as responding to prompts sent not this time but in previous transmissions.
Is this problem occurring only for me?
I don't think this is a problem I can solve on my own, so at the very least, I'd like to know if this is a phenomenon affecting all users.

Incidentally, I did a data export and checked the chat history, but even there the answers were interrupted and memory clouding occurred.

[hidden night]

Mine is (I think) doing the same things as yours (couldn’t find original post), where it just, stops writing and deletes the whole message

[astral vessel]

Hey guys , really need help here. I am stuck 😦

[fiery eagle]

Where can we see Sora prompts? That's great 🙂

[onyx kraken]

Prompts are below the vids, and can scroll to the side to see more vids. https://openai.com/sora

[fiery eagle]

thank u

[craggy topaz]

Hello. For the past few days I have been trying to do my first fine tune but I keep getting this message. I don't know if it's an OpenAI error or something wrong with my file. I am Tier 3, have high limits, no usage yet this month. Any ideas, please?

[grim urchin]

Can i access sora

[charred parcel]

no, it's not publicly released

[bright hare]

is there a setting to make chatgpt's responses appear all at once instead of word by word?

[past geyser]

Only by editing CSS with something like a browser extension. I feel like it's a good option and easy to implement though, so maybe pop it in #1070006151938314300 ?

[brave crow]

experiencing extreemly slow responses: 1 word arrives in 2-3 seconds. Is that common for long chat discussions with long text in each message?
Also, chatgpt got confused, started repeating itself, the same paragraph multiple times appearing in responses.
Do you suggest I have shorter sessions?

[slate geyser]

Hi all can we get early access to SORA ?

[rotund radish]

Howdy, qq, I have a GPT doing a custom action, its doing a simple api call, then my api needs to go and do stuff. How do I post back to that GPT? there's not a response url that i can find

[trail cairn]

no

[split gazelle]

Hi All I am seeing the above error for past few days. Tried contacting openai help via email but no update yet. Can anybody help me with this?

[rocky rover]

Report this to #1070006915414900886 and follow the template OAI asks for on how to reproduce this. OAI staff is active looking these reports.

[split gazelle]

Thank you I will do that now.

[versed dagger]

Hey developers 😄 !
I have a question: Do you think it's smart to implement a lot of functions to an assistant from assistant api? Cause i want him to be able to do product recommendations based on user queries, that can be complex, for example "give me something under 5$ that i can do ... with", so the assistant would need to filter the data first and then decide based on it's knowledge or "intelligence".

[supple agate]

Hey all, I'm hoping to speak to someone who has had some experience with the Assistants API, each time I try and do something I hit a problem... examples are:

1. I'm trying to build a travel assistant, the Terms and Conditions is a file uploaded, the assistant randomly starts referring to it.
2. The assistant tries to find out how many travellers and to determine how many are children and how many are adults.. It simply does not get the concept of going back to the user to ask for further information
3. The assistant will regularly go off and start asking about budget even when budget hasn't been mentioned...

It'll all very random and I"m getting close to the point where I just hand cut the code that manages all of this but wanted to check I hadn't missed a trick....

Happy to jump on a call if it helps.

[brisk pendant]

Hello! If I update GPT's instructions, do I need to make new chat for them to take effect?

[timber shadow]

Hi, yes you do! Chats started with the old version will stay locked on that old version, they don't update on the fly. There will be a message in such old chats at the bottom telling you a newer version of the GPT exists, too.

[brisk pendant]

ah, thank you :)

[craggy topaz]

Can anybody help with this error when I try to create a finetune? It says I exceeded my quota but that can't be it.

anyone run into this beforE?

[brisk pendant]

how does the ChatGPT teams work? Like I invite someone to my "team" and they'll get the premium benefits too or?

[timber shadow]

Yes, that's pretty much it! Team subscriptions have a minimum requirement of two seats. Each seat has its own profile and usage cap, and the Team can share GPTs within the group.

[timber shadow]

Hello! Is all your billing info set up right? Do you have credit available on your account? You might try in #dev-chat too!

[brisk pendant]

but do they need to login with my email or can they use their own existing account xd

[timber shadow]

They have their own login and account, what I think happens is you send them an invite link, then they use their existing ChatGPT account to join your team.

You won't have access to each other's chats, accounts, etc.

[brisk pendant]

okay thanks again :)

[timber shadow]

Any time!

[molten portal]

Hello, does anyone else have the same issue? I've tried different browsers, but GPT-4 isn't working. Tech support is silent

[timber shadow]

Hello, can you try in a private window with all browser extensions disabled?

[molten portal]

Tried, nothing changes, 3.5 works, 4 doesn't

[timber shadow]

What browser(s) is it happening in?

[molten portal]

In any browser, tried FF, Chrome, DuckDuckGo.

[timber shadow]

Do you have a VPN installed?

[molten portal]

yes

[timber shadow]

What VPN do you have?

[molten portal]

I use my own VPN server on Hetzner.

[timber shadow]

I see -- have you tested with the VPN disabled to see if it's any different?

[molten portal]

Yeah, it's about my server, it worked without a VPN. Is there a way to unblock my IP?

[timber shadow]

I'm not sure -- based on what you're finding in the Network tab though, do you think you'd be able to whitelist anything on your VPN? Or does it not appear to be a filter thing, but more of a block thing?

[molten portal]

On the server, there are no blacklists or whitelists; the issue started after the crash on the 14th. 😦

[timber shadow]

Maybe try submitting a support ticket at help.openai.com. Open the chat widget in the bottom-right, click through the most relevant options, then you'll eventually find a path that lets you type your own message. This will submit a ticket that OpenAI support can review and get back to you on. It may take some time. You'll get an email when they reply to you!

[austere prism]

As of yesterday, after the issue being fixed for less than a few days, I am once again receiving the conversation key not found error when sending a second message in conversation with chat history off. This has been going on and off now for months. I have reached out to support and they only responded with templates and didn't actually read what I've said. Can some please help actually address this once and for all? How is this so difficult? Surely you can proactively monitor on the client side for errors so that you are at least more aware of this problem and can find a more permanent solution. It's a bit silly that I can essentially only use chat history off for a single response each time, that's not productive.

[timber shadow]

The chat history/training off bug that causes conversation not found errors was acknowledged by an OpenAI team member earlier this month: #1203014345106395156 message
So hopefully they're able to address it soon.

[burnt grotto]

Hello all! Anyone experienced with the API and willing to help?

[timber shadow]

Hey! Are you having some issue in particular? Also, the folks in #dev-chat might be able to help ya too.

[small gull]

I too .. am here for the Sora

[austere prism]

Indeed, a thread I had also participated in. I just can't fathom how after months they still haven't found a more permanent solution. Should this be flagged to the same employee again given that the thread has closed?

[twilit jolt]

Hi , does anyone know if there is an Open AI newsletter one can sign up to, please?

[timber shadow]

I don't think that would be necessary personally -- sounds like they're aware of it and are working on it. If it continues to persist, you could try making a new post in #1070006915414900886, just be sure to follow the pinned report template. I don't think a thread being closed necessarily means the bug has been fixed, I think perhaps that thread was just getting off topic.

[craggy topaz]

absolutely, billing is fine, limit is high, I am tier 3. I will try the other channel as well, thanks

[timber shadow]

Hi again! Can you try in a private window with all browser extensions disabled?

[normal pebble]

All of my history, previous conversations are gone, and ChatGPT is gone

[normal pebble]

I'll try

[timber shadow]

As for history: see that little gray arrow on the left side of your screenshot, right in the middle? Click that, your history sidebar should reappear.

[timber shadow]

Is that happening when you try to click into any chat?

[normal pebble]

Yes...

[timber shadow]

And is starting a new chat still not working either? Have you tried in a private/incognito window, with all extensions disabled?

[normal pebble]

It works in an incognito window

[timber shadow]

I see, it may be a browser extension interfering then. You might try turning all extensions off, making sure it works in a normal window, then turn them on one at a time to see which one might be interfering, testing after turning each one back on.

[normal pebble]

1.) I only have AdBlock - The Best Adblocker on. And usually, it works 100% fine
2.) I paused it this time, and it STILL doesn't work

[timber shadow]

Can you try disabling the extension completely, not just pausing it? Then clear your browser cache, and close/reopen your browser, then try again?

If it still doesn't work, you might check your security settings of your browser. There might be a browser setting interfering as well that's active in a normal window but not in an incognito window.

[normal pebble]

I deleted the AdBlock, and cleared an hour of my history... It's working fine now.

But I still don't understand. AdBlock was open whenever I used ChatGPT, and they worked just fine...

[rancid pivot]

Has anyone had any issues with uploading .csv files to custom GPTs? I am trying to make a GPT to read .csv files but it doesn't seem to recognize the file whenever I upload it.

[ashen spindle]

When I want to add a payment method in Openai, I get the following error

[wary wraith]

Hello, everyone!
I have a question, in the company where I work, I would like to generate API keys for each client and track the API spend of each client separately, is this possible?
For example, I have customer X and customer Y and both have different keys linked to my OpenAI account, when both spend, I can see the spending of customer X and customer Y
I would also like this to guarantee the security of the data sent via the API.
Would this be via the Enterprise API?

[elfin scarab]

Is anyone well versed with NLP? I am working on an university project and need help but its kinda urgent. So please contact me. I'm a newbie but I think might be simple for a decent person from the domain.

[rustic schooner]

How might I apply, as a career professional artist, and now AI artist, to be a tester for SORA?

[timber shadow]

There's no application process in place currently. Current testers are red teamers and some creators they invited. No news about potential expanded access yet!

[burnt grotto]

Specifically, looking to upload a document using the Assistants API, but the documentation I found was for fine tuning, so not sure it is applicable. Essentially, I want to upload a document so that I can ask GPT to retrieve information from it. Thanks a lot! I did ask in the community-api-help, will try api-discussions too!

[timber shadow]

These pages are related to the Assistants API, perhaps they'll be helpful if you haven't already seen them:
https://platform.openai.com/docs/assistants/overview
https://platform.openai.com/docs/assistants/tools

[final zealot]

Hi, I've been a while away from reading daily themes... Im back and looking for daily credits. No more 4 credits per 24hrs?

[burnt grotto]

I have seen them, but looking specifically for how to upload documents through the API. My understanding is Assistants API is where it is possible, but if it is through ChatCompletions that works too

[timber shadow]

I think that might be found in the "Knowledge retrieval" section in the Tools links above:
https://platform.openai.com/docs/assistants/tools/knowledge-retrieval

[burnt grotto]

(same way you would upload a document in the chat interface)

[timber shadow]

Ahh that would probably be this section: https://platform.openai.com/docs/assistants/tools/passing-files-to-code-interpreter

[timber shadow]

Hello! What happens when you try the /daily command in #daily-theme? I honestly have never tried it myself, but you can generate 5 DALL·E 3 images a day over in #image-bot! Just start a message with /draw over there to get started.

[burnt grotto]

even if files are not code files? thanks a lot for helping 🙂

[timber shadow]

It does not need to be code files, no. Here's a list of supported file types: https://platform.openai.com/docs/assistants/tools/supported-files
You are welcome!

[final zealot]

Thanks but it might be not enough. I used them for vertical or horizontal overpaitings images generated via Bing creator with D-E3

[timber shadow]

Ahh on Labs? I am not sure about that -- is the /daily command no longer working for you? I think there are still 15 monthly credits issued on Labs.

[final zealot]

nope. Cant use it / doesnt works. And Yes credits for Labs ;] Sorry not to mention. Anywa thx. I dont think thera are 15 monthly credits for free users. Im overpainting my pics only occasionally, so didn't noticed they are gone 😉 untl I need them.

[timber shadow]

I've never paid for Labs and this URL shows 15 free credits for me, so maybe double check this? https://labs.openai.com/account

[final zealot]

These 15 are for a start, they are expire after 30 day if You won't use them. Obviously I dont have them now 😉 coz I logged on Labs 5 months ago or so (and used 4 free daily credits from daily-theme channel). Something must changed. So not bother You more . Thanks for Your support.

[ashen spindle]

Ok, I tried a different card, I received an SMS, I entered and it rejected my card.

I will use it for API

[karmic perch]

Does anyone know of any free tool to visualize embeddings in 3D?

[lunar slate]

can someone whos experienced with OpenAPI and creating API's help me? Im making a gpt that makes use of an API (my own API), and i thought I set it up properly but it isn't working as intended.

[astral vessel]

Hi guys, I am facing this issue, which custom-GPT wont connect with my server.
Which I use full link directly in browser it works.
Error from custom-GPT config and complete server URL given
Is the URL in path wrong or what, I stuck 😦

[daring imp]

guys i have a question on how to access sora. May you can help me?

[rugged wing]

i was using chat gpt for a little project of mine and this was its response ill make a video now but ill just send what it said first

@me or reply with help

i think i broke it or found a bug

[solid reef]

Has anyone received the ability to call in @-gpts yet

feel like I'm a year behind everyone else receiving updates

[rustic marlin]

AFAIK, everyone has it

im pretty sure that calling custom GPTs with @ wasn't a partial release and everyone got it at once

[solid reef]

I've been typing @-gpt-name on various browsers for a while. reinstalls and clearing caches etc

[rustic marlin]

when it happens that the AI starts to repeat itself, repplying further is fruitless as it is already primed to keep doing that, it is best to just edit the message and try again or start a new chat

[rugged wing]

ok thanks

[timber shadow]

Earlier this month Eric from the OpenAI team mentioned that the rollout wasn't at 100% yet, it's possible that's still the case. #1203945630561599518 message

[rustic marlin]

oh, so it is a partial rollout then

good to know

[heavy cape]

Anybody know how to get access to Sora and if there is a waitlist to join?

[red lance]

@tribal talon are you here like I need to ask about something related with ai so can you help?

[ionic quest]

How can I turn a vector embedding into cartesian (2D) co-ordinates?

[tribal helm]

Sora is not working. It showed me the video generated 18 hours ago by Open Ai. Maybe they are working on it?

[rustic marlin]

there is no way to access it currently

[zenith solstice]

Is ChatGPT updated tpday?

[zenith solstice]

I just found it's me diffifult for extracing documents

When I ask more details it just refuse to go through the document

[cold folio]

is the file too big? whats the file extension?

[zenith solstice]

I'm pretty sure I used the same size/kind of file as yesterday.

And when I ask for details, it just refuse to search to the document

And I was normal until yesterday

btw I'm using "Document Assistant"by ChatGPT

[wraith valley]

Hi, does anyone have problems with ChatGPT? For example, everything you ask them says they can't do it, I have ChatGPT Plus

ChatGPT says it can't browse the internet, it can't analyze websites, it can't write texts, I ask it for all this and it says it can't, I have to tell it several times that it can until it does, but that's how I spend all the tokens I have. I have and then I have to wait until I can use ChatGPT Plus again, please fix this

This is the message from ChatGPT Plus:

I understand your frustration and appreciate your patience. Unfortunately, I cannot directly fulfill the request as presented due to current policies and limitations. I am here to offer guidance, advice and assistance within the permitted capabilities. If there is anything else I can help you with or any other questions you have, I will be more than willing to do so if possible.

[onyx kraken]

Try starting a new conversation.

Also, if you're willng, what's the actual request you send to the model? What's your first prompt in the new conversation?

You can get what I call a 'magic new chat window' where things go weird. Sometimes good, sometimes bad.

I would not waste a second message in a conversation where the model wasn't acting 'right', because it's likely to continue to act 'weird'. If I don't like the weirdness, I won't put another message into it, I start over.

[wraith valley]

I already tried opening several chats and in all of them the same thing, my first prompt is that I be an expert in copywriting and Google Ads

[wraith valley]

For example, for it to do what I ask, I have to spend 3 or 4 prompts for it to do it, because it sends me the message I passed above.

[somber void]

Hi there,

I verified a domain and added billing details, but can not complete verification for some reason.

Does anyone know how to fix this?

Context: I went through this process before but decided to update the link and am now stuck in this verification stage. The toggle switches (name and links) are blocked now

Also, all my 30+ GPTs are unpublished now 😱

[onyx kraken]

If you are literally giving this as your first prompt:

"I be an expert in copywriting and Google Ads"

A problem is that is not an instruction to the AI.

It does not know what you want it to do.

It will instead discuss or congratulate you on your success of being that expert, or ask you questions to show it admires and values such an expert, or otherwise reacts - but you haven't told it what you want it to do, so it isn't doing it.

[rugged wing]

This has been happening since yesterday with chat gpt4, no vpns installed, no plugins. I tried it on my phone and it doesnt work there either.

it only happens with images

[peak kelp]

Yep I can’t upload images either since two days?

[onyx kraken]

Upload and discuss images has been working fine for me, not sure how to help, but I can't duplicate the problem.

[peak kelp]

it’s always like this

It does keeps loading for 5 - 10sec - then this error shows

[lunar slate]

Not accessible to the public atm

[timber shadow]

Does it only happen in the app version, or also the web version?

[peak kelp]

Both

[cold folio]

Have you reached the message limit?

[peak kelp]

Nooo

[scarlet silo]

I'm having the exact same issue and would be interested to know too.

[rugged wing]

its your account, I had to rebuy the plan and create a whole new account and now its working fine :((

[gloomy quail]

Can anyone help me on using sora? I am now a gpt4 user

[foggy veldt]

Hello! I'm not sure if this is the channel to ask this, but I'm a Ssr QA Tester and also have a degree in film-making. I'm currently working on some personal film projects and want to learn about Sora. Could someone contact me or guide me on how to help with the testing of it?
Thanks! By the way I'm from Argentina and I'm a big fan of Open AI ❤️ been using it since the beginning 😌
Thanks everyone!

[slim cradle]

Has the uploading image been fixed yet? I'm still not able to do it

[median raven]

Can anyone recommnd ai jobs/internships online? would be a great help

[golden ruin]

hi there, we're not using this channel for job opportunities 😊 although you can ask about anything OpenAI related

[median raven]

Fine, let me rephrase, Openai internships (this is openai related)

[charred parcel]

OpenAI opportunities are listed here https://openai.com/careers

[peak kelp]

still doesnt work for the image upload, someone else have that problem too?

[unborn axle]

hello everyone i have my product photo of spine massager i want ai to generate 3d image of 3d video to demonstrate or person sleeping on the massage bed the person i want to generate ai based can any1 help me with prompt ?

[weak ruin]

Why is the turbo so slow in compare with 3.5 turbo

[orchid heron]

HELP?! I am hoping someone else has run into this and knows how to resolve.

I was using ChatGPT plus for ages. Then I upgraded to ChatGPT Teams plan. The teams plan did not work for us for a variety of reasons so I CANCELLED my teams plan (there was no option to downgrade). My intention/goal is to go back to ChatGPT Plus. Now, when I try to access ChatGPT I am locked out with a message that just says, "workspace deactivated". Anyone have ideas how I can get back to GPTPlus?

[golden ruin]

it will be better if you reach out to their support directly from https://help.openai.com

[weak ruin]

playground does not not work

I mean the assistants

I tried the chat version now but the code is not formatted correctly in the respond

[trail cairn]

playground works for the purpose its made for, api testing. its not supposed to be chatgpt replica

[tired nymph]

is there an error again with chatgpt?

[weak ruin]

okay but do you know why it's being lazy. I can't get any good code without saying rest of your code here

[waxen ice]

what timezone is the deadline for superalignment grants? AOE or PT ?

[wraith valley]

ChatGPT Plus is not working, Open AI, solve this quickly, I ask ChatGPT for things and it just says it can't do it, it's the only answer left, I don't understand why it takes so many days to be able to solve this problem that I see you are having the most

[onyx kraken]

Be sure you start a new chat.

If 1 conversation is saying 'no', maybe that is a bad seed. Start a new one.

If you have trouble in a new one, share what you are asking the model to do.

Or check here and here: https://openai.com/policies/usage-policies https://openai.com/policies/terms-of-use

Some things if you ask the AI to do, it will refuse because it must.

[static cloak]

Hello All I am GPT-4 user, I was just wondering what is the difference between gpt-4 -turbo and chat gpt 4.

[timber shadow]

Hello! The version of GPT-4 on ChatGPT Plus and Team is a 32K version of Turbo, so they're the same, just not full 128K Turbo like is available via API and ChatGPT Enterprise.

[static cloak]

ohh, also is there paid version for codex

like I saw in a youtube video "Creating a Space Game with OpenAI Codex"

how can I do that?

[timber shadow]

Codex was deprecated earlier in 2023: https://platform.openai.com/docs/deprecations/2023-03-20-codex-models

Although I think GitHub's Copilot might use an implementation of Codex, you might try asking in #ai-discussions. According to the deprecation notes on the link above, OpenAI recommends GPT-4 as a replacement for Codex.

[static cloak]

thank you I didn't know which channel to ask this lol, I will ask in AI discussions.

[timber shadow]

#ai-discussions is just a good place to talk about non-OpenAI services like GitHub's Copilot. You're in the right place for the questions you asked! Codex is just deprecated.

[orchid heron]

Thank you - I did this as well as messaging here.

[ocean lake]

Hey.... not sure if this is the right place for it. I'm trying to add credits to my account. I've got all my card info correct, There is more than enough funds in the account, but for some reason when I click "add card" instead of verifying my card, it says "card declined" and won't let me proceed... Is this a known bug?

I just got off the phone with my bank, they can see the verification attempts and the acceptances, but it's not adding it for some reason

[ionic grove]

having trouble browsing direct urls, any suggestions?

[ionic grove]

What happened to the messages

[compact plume]

Hi I am working on a Custom GPT. I uploaded documents to the knowledge base. However, when I start a new conversation and the model begins to Search, it suddenly starts sending a reply saying it has no information about the topic. Meanwhile, the "Searching the knowledge" message still appears actively above. Any advice? I've tried custom instructions. Code Interpreter is the only capability I have turned on.

Like, it's still searching yet it answers anyway. Is it a bug or something?

[onyx kraken]

There is a new report from a few users about custom GPTs that were working fine suddenly not acting right/knowing their knowledge. https://discord.com/channels/974519864045756446/1208676441101963295
#1208716921470713916 message

I'd encourage you to add your report to the bug report if you're willing

[compact plume]

Done. I hope it's just a bug that can be fixed easily.

[slender island]

Guys i need some help ,its becoming so annoyingggg

i am using gpt 3.5 the free tier

going on i write a prompt to chat gpt it can be a math problem or a simple quesstion

once chatty finishes the answer it rewrites again the answer multiple times

[onyx kraken]

I've rarely seen that. Have you tried to just start a new converstation and see if that fixes it?

[halcyon terrace]

Anyone else unable to upload docs to GPT4?

[dim sigil]

The other problem is ChatGPT and CustomGPTs refusing to read the documentation then proceeding to tell the user to do so instead... Borderline useless response.

[onyx kraken]

New bug? Been happening always?

[ionic grove]

Back up your GPTS's!

[dim sigil]

New-ish. I've been getting it as a response significantly more in the past few weeks.

Mind you this is a response for coding prompts

[onyx kraken]

Great idea to report it, if you haven't yet. #1070006915414900886 if you wish

[frosty basin]

Hey guys, I have a problem with GPT 4.0

Everything works normally except when I upload a photo, any photo for that matter, and it just gives me this error. Messaging it works normally

I tried on another device, different network, cleared my cache, different browser, used incognito, logging out and logging in, basically everything yet it's still occuring. It's even happening on ChatGPT's iPhone app. This issue is only in my account, my friend with a different account does not have this issue.

Does anybody have any clue on how to fix this?

[dim sigil]

Is it the same image every time you tried it? If so, it could be a corrupted/improperly formatted file or it's rejected by the image handler by openai for some reason (Labeled inappropriate or whatnot)

[frosty basin]

I've tried several photos, it just doesn't want to work. I'm using PNGs. Previously, it worked with the screenshots I'm taking via my macbook

[vivid mortar]

I'd say maybe logout and log back in, you said you've tried multiple browsers?

[slim cradle]

Same issue has been happening to me. I’ve tried different devices, browsers, images, logging in and out. Nothing seems to work

[frosty basin]

Yup, that did not help

Support isn't that helpful too, they reply with a generic response than don't follow up. I'm having to make new chats every time. Someone mentioned escalating the situation, however they did not email me with anything, and did not reply when I asked rfor an ETA

[scarlet silo]

Hi there,

[marsh cobalt]

since yesterday I have this problem and I dont know why

the prompt with picture doesnt work

[frosty basin]

same here, exact same problem

i've had it since friday

Just talked with OpenAI support, it'll be "escalated". Not sure how accurate that is, but I'll keep you guys posted

[nocturne lodge]

anyone facing the same issue? gpt not being able to read images at all. been facing this issue since friday. thought it was a bug but has not been fixed yet.. with other images too not just this i tried with other images.

[marsh cobalt]

I hope that open AI is going to reward us with something because of this problem...

[onyx kraken]

I've seen a few people talk about this in the last few days. I personally am not affected.

I recommend checking out #1070006915414900886 and making a report, or adding your details to someone else's if someone else has reported it (I have not reported it because I am not affected at all, I have nothing to report).

[nocturne lodge]

ill do that, thanks for the heads up!

[wanton wolf]

guys i need help, this error keeps popping all the time "Our systems have detected unusual activity from your system. Please try again later." I cant use my chtgpt plus more than 30-40 minutes and this starts coming up on my chat. i have to open a new chat again to continue

[nocturne lodge]

i guess many other fellow users have pointed it out aswell, hopefully it is taken care of soon enough..! thanks once again. have a nice day :DD

[onyx kraken]

If a new chat works, that makes me wonder if you can soft-fix it by refreshing your browser window. Does that help?

[wanton wolf]

nope i tried everything including logging out and back in

i cant use the same chat with the old data but i can start a new chat and that works

[onyx kraken]

Hey, thanks for reporting it! OAI seems to have an easier time keeping track of the bugs there than the endless multi-topic chatter we make in these threads here 🙂

I hope they get it fixed for you soon!

[onyx kraken]

Fascinating. You're the first I've heard report that pattern!

Please do #1070006915414900886 and explain it, perhaps provide the url string from an affected chat (that looks like this: c/cdc8380d-5189-4768-8b60-e63751a65fce and their engineers can use that to check stuff out, may or may not be needed)

[wanton wolf]

cool i will do that

[molten bane]

Hi, The ChatGPT4 (Plus) got really slow in typing recently. It takes around 3 minutes to type one response, something that ChatGPT4 previously accomplished in 10 seconds. Now, it takes a very long time.

I am using the Chrome browser with the latest update as of today.

Any advice is appreciated.

[soft moat]

Same here

[quasi mist]

Ah, having this problem on Microsoft Copilot as well

[sullen raft]

guys

suddenly I am getting weird chat histories

[past geyser]

Important that you immediately change your password at the very least. Your account is likely compromised.

[sullen raft]

I just changed my password

was getting weird GPTs top left in mandarin

how can you get additional support to be safe

[past geyser]

a strong password should keep you safe or you can login with a 2FA Google account I believe

[sullen raft]

how do i set up 2FA

[past geyser]

You can't if you have an existing OpenAI account registered by email. If not or if you're happy to start a new account, use Google to sign in and use their two factor login

[slender island]

i tried thousends of times it continued to do the same thing

but on my phone it worked fine

[bitter valve]

Hi, does anyone lose access to Dall E3? It has been a couple days since that option gone from my chatGPT

[trail cairn]

what option? dall-3 has not had its own dropdown selection for several months now

[royal adder]

Hi. I encountered an issue with chatgpt4. chatgpt 4(web) not sending messages despite chatgpt3.5 working and subscription active. Just after hitting enter the message is send but the new chat is not created. Attached screens: 1) chatgpt3.5 2) chatgpt4. Had this issue for several days. Switching devices/browsers, going private didnt help

[trail cairn]

what are the errors in the console view since it shows 7 and 8

[royal adder]

on message send script fail to load

[trail cairn]

what about the other red errors? or were those unrelated to chatgpt

[royal adder]

they seem to be from multiple attempt to send message, the same error that is encoutered on send

[trail cairn]

if something alters the script or blocks it loading fully it could cause that but you said you tried on another device?

[full surge]

hi guys i need a hand, please,
thanks again for such boring problem
I have the other error , when i check the saved conversations i got the error
Unable to load conversation d0d9cf13-58be-4fd1-b702-405aa

[trail cairn]

that error in the pic can be caused by different vpn/antivir/browser privacy services, extensions etc.

[full surge]

thanks dear Radium i will try bro.

[royal adder]

already tried it on another browser and devices didnt help

why chagpt3.5 working then

[trail cairn]

another browser or even a device wont help if the blocker is for example a vpn software or a blocker dns service

the arkose error is probably arkose labs script that is being killed

[tall sky]

Hey folks, my ChatGPT Vision stopped working, every time I upload a picture I get a server error

The pictures always upload correctly, then when I press enter ChatGPT "thinks" for a few seconds and invariably it returns a server error. It's been going on for a few days and it's super annoying, anybody else experienced the same? Happens both on mobile and desktop, tried to clear the cache and reinstall the app to no avail

Support seems to be useless, I got just an automated reply via chat on OpenAI website, than nothing at all when I gave more details

[timber shadow]

Hey, there are some other users in this thread reporting the same issue: https://discord.com/channels/974519864045756446/1208813258086027364
Maybe add your experience there! OpenAI team members are in that channel fairly frequently too. As for the support chat on help.openai.com: it's not currently possible to connect to live support, but if you got through the default options to where you were typing your own message, that submitted a ticket that they can review and get back to you on, it may take some time. You'll get an email when they reply to you!

[tall sky]

thanks man! will look into that thread right now

[fathom roost]

i dont know why but i keep receiving this error continously for the last 3 days. It only does it when i send an image tho...

[fathom roost]

same issue!

[onyx kraken]

Interesting! I recommend making a #1070006915414900886 about it and including at least 1 shared conversation showing the bug in the post you make in bug reports. OpenAI employees (gold names) sometimes even directly discuss issues in that part of the Discord, it's much more likely you'll get someone who can actually help if you give them the information you can.

[granite bobcat]

Is it possible to integrate a **custom GPT **chatbot into my website? Using Remix, Tailwind CSS, and React with TypeScript. I've added a "normal" Chat GPT but having trouble understanding if you can integrate your custom gpt.

[timber shadow]

Hi! Custom GPTs live only on ChatGPT. To create a GPT experience on your website, look into the Assistants API -- that's the API correlate to GPTs on ChatGPT. https://platform.openai.com/docs/assistants/overview/agents

[drowsy fjord]

Does anybody know about data residency? If I am a cooperate customer can I choose that all data will stay in europe? I am asking for regulatory reasons.

[tacit wing]

My custom GPT now started using Python code to look into PDFs, that is new - and it's giving me a worse result:

Is there a way to get last week's behaviour back? :)

[fringe heath]

huh

[tulip egret]

yes same issue with images only, neither able to give it as a prompt nor able to download it from them

This bug is stopping me from doing anything.

[scenic spear]

how can we fix that bug, anyone please?

[fathom roost]

Hello,
​
Thank you for reaching out to OpenAI support.

We appreciate you bringing this concern to our attention and want to assure you that we are here to provide further assistance.

Sorry about the trouble. Sometimes, users of ChatGPT might see errors like 'Something went wrong' or similar. The error you've experienced could be due to a range of factors. It might be an issue on our end or possibly a problem with your internet connection.

Here are some steps you can take to try and resolve the issue:

1. Retry the request again later - this often solves the problem.

2. Visit our status page to check for any active outages that might be temporarily impacting the service.

3. Clear your browser's cache and cookies.

4. If you're using a VPN, try disabling it.

5. Try using an incognito browser or a different internet browser/computer. Sometimes, a security add-in or extension can cause this type of error.

6. Disable any cookie blockers you might have enabled.

7. Consider connecting to a different network, such as switching from WiFi to a cellular/mobile network

Even without a reported outage or issues with your network, these errors can randomly occur. While they're less common for ChatGPT Plus subscribers, they can still occur from time to time. In most cases, retrying the request later resolves the problem.

We hope this helps! Let us know if there's anything else we can assist you with.

Best,
Keboy
OpenAI Support

the mail i had but i tried all of those

[charred parcel]

what is the issue you are facing?

[fathom roost]

i dont know why but i keep receiving this error continously for the last 3 days. It only does it when i send an image tho...

[charred parcel]

ok, I am getting the same error. I don't know why.

[charred parcel]

have you tried on the iOS/Android app?

[green patrol]

#Privacy: Does openAI still train on deleted posts? Please answer with a source quote. (I know about switching off training for 30days backlog, but this is a different question)

[elfin gyro]

Anyone here have experience getting actions to work?

I'm attempting to make a gpt personal assistant which can access my calendar and emails through zapier, but keep getting an error that says: "Error talking to"

Joined this community bc I couldn't figure it out online

Also if anyone knows of something like this that already exists lmk

[sullen raft]

Guys

So I was thinking maybe my account got compromised

I changed my password but I don’t know if it fixed anything

I don’t see any new chat history but these new GPTs keep getting added to my GPT list and they’re all in mandarin

I keep removing them but they keep coming back. How can I get additional support on this pls?

[fathom roost]

i did! same error

[sullen raft]

How do I get additional support???

[past geyser]

Sorry, I found a support article but it was specific to API keys. If you've changed your password it's the most important step, but have you also changed the password of the email account it's tied to? That could also be compromised. The odd GPTs popping up could be a hangover from when the account was compromised but it's best to play it safe. Are you able to completely close the account and start a new one?

[past geyser]

https://help.openai.com/en/articles/8485685-my-chatgpt-account-was-hacked-i-m-seeing-unrecognized-conversations

This one is better for you.

(and yet still no new information beyond "Change your password")

[sullen raft]

I changed my password immediately but these GPTs in mandarin keep popping up

[sullen raft]

I pay for gpt plus I haven’t thought about closing it yet

Let me update the email password too

[past geyser]

Again that could be automatic based on prior use I believe

Here's how you can report it for investigation:

https://help.openai.com/en/articles/7242626-how-can-i-report-fraud-or-suspicious-activity

[sullen raft]

I don’t know the mandarin language also never used any prior GPTs in the top left. Maybe those could be automatic since I asked questions regarding engineering topics but idk

Just want to play it safe since this has never happened until yesterday

[past geyser]

Probably more likely because that language has been used historically by the account

[sullen raft]

By my account? I only message in English

[sullen raft]

Where do you submit for investigation then?

[past geyser]

See the last link I provided. Use the support chatbot

[onyx kraken]

Hey, I wanted to ask about the idea of completely closing an account and starting a new one.

We appear to have a limited number of times a mobile, not land-line phone number can be used to create a OAI account, before we can't create new accounts with that number.

The same limitation with the same email, though it's reasonably cheap to get a new email - but it can be expensive and difficult to get a new mobile phone number, plus one that has never been used to make an OpenAI account (those numbers get recycled as users give them up, so I imagine a day when there's just no new mobile phone numbers never used that I can get! And it's not like I can ask the cell phone carrier to check for me, I don't believe they do that or easily can check if it's ever been used for an OAI account).

So, to me 'close account, create a new account' sounds like an absolutely desperate, dire, likely to be a real problem nuclear answer to use - especially since closed accounts can never be reopened if I understand the helpfiles correctly.

At best, I'd consider 'create a new account, make sure I actually can with my available resources, then if I'm completely happy there, close the old one' - but that's also a nuclear option. There's only so many times any of us can do that, before we're not going to have a acceptable phone number available.

Or, is it not as bad as I understand, is this actually easy and certain?

[past geyser]

This user in particular has previously enquired about 2FA which is only roundabout available, through authentication with a Google-managed account. For some additional security measures this is the way to go and in doing so there's no arbitrary limitation on how many times you can connect a phone number. That'd probably be my advice but I understand the disincentive is existing chat histories, settings & payments. It's an option though.

[onyx kraken]

Thank you!

[novel valley]

Hi guys the "send message: button is not working

[sinful mirage]

It's working fine. Did you try in both web and app ?

[wispy olive]

Hello, how to fix it?

[timber shadow]

Hello, does it happen with every prompt, or just some?

[wispy olive]

I uploaded my character, gave a description, asked him to generate it in a realistic style and he gives me this time after time

[timber shadow]

I see. As a test, can you try a very simple prompt in a brand new chat? Something like "Please create an image of an apple." I'm just curious if it could be a content policy thing, and not a bug, but I'm not sure what your character might look like, or what your description is.

[timber shadow]

Maybe try in a new chat, it seems like it might just be a temporary bug.

[wispy olive]

Alas, I tried it) I even used drawn to style

[sullen raft]

thanks boss

[jagged shale]

What is the best practice to prevent something like a "DDOS" Attack on API assistants?
A user could just create a python script and spam our assistant (deployed via chatbot on a website) and spend 100$+ within a few minutes.

[past geyser]

That's probably a question better suited to #dev-chat but it's also not necessarily a problem unique to OpenAI GPTs - there are lots of ways to rate limit users depending on your front-end. Some wider research might be required.

[modest jetty]

Anyone here knows how to remove the usage cap in the GPT4 plan?

[vivid mortar]

There is no way to remove the limit, it is a set limit based on usage during a timeframe

[barren pine]

Built a custom GPT, but I want my friends to be able to use it without paying for a premium subscription. Wondering if there was a site that allows me to take the prompt and an open AI API key and upload it and allow my friends to use it?

[onyx kraken]

Be aware, this is what OpenAI says about 'sharing' API keys: https://help.openai.com/en/articles/5008148-can-i-share-my-api-key-with-my-teammate-coworker

[onyx kraken]

As far as Custom GPTs vs API... I don't know, but based on this help:

https://help.openai.com/en/articles/8798878-building-and-publishing-a-gpt

"Will my GPT be available to all Plus users when published to the store?
Yes, published GPTs will be available to all Plus users. Free users will be able to see your GPT, but will not be able to start conversations with it."

It appears that it is not intended for custom GPTs to be usable except by those with ChatGPT+.

However, you can possibly use the Playground and through API access directly create something the rough equivalent of a custom GPT, which can then be accessed through a website or other ways.

Here's a forum post about that: https://community.openai.com/t/how-can-i-use-my-custom-gpt-on-my-website/528694

[topaz cobalt]

Hey guys, i've made an assistant on openai's API platform. How can i make my Python code access the specific assistant i've created on the website?

[thorny belfry]

Hey. I'm experiencing an issue with ChatGPT 4 on both my personal account and my ChatGPT Team account, both of which have paid subscriptions. After sending a message, it simply disappears, and nothing happens. ChatGPT 3.5, on the other hand, works fine. Is this issue specific to my accounts?

[naive jasper]

where can i use sora

[topaz cobalt]

have you tried on other browsers or accounts? Had the same issue on chrome, but not on edge or Opera

[granite bobcat]

Hey @timber shadow would that mean that the custom GPT:s could be used on your own website through assistans api or am I misunderstanding?

[thorny belfry]

I just tried using Safari instead of Chrome and encountered an issue where I can't log in. The 'Continue' button seems to be unresponsive... weird.

[topaz cobalt]

seems like a bug then, maybe you should report it to the bug-reports channel. Have you tried to see if you can login on the openai app?

[rustic marlin]

to be sure it is not cache, try to use an incognito tab

[thorny belfry]

I've tried using the 'Report a Bug or Error' feature in the chat, but unfortunately, I received no response. And I can't keep it open forever. Is it possible to submit a report via email? I couldn't find any email to write

[thorny belfry]

Tried, the same

[rustic marlin]

how long have been since it started happening?

[thorny belfry]

I haven't used ChatGPT for a week until today. Today is the first day I've encountered such issues

[rustic marlin]

this sort of problem is usually something that gets solved with time, try to just wait a bit

but if it continues, you might need to reach the support

[charred parcel]

best to ask in #dev-chat

[topaz cobalt]

oh mb, sorry

[zealous valley]

Hey everyone, I am on ChatGPT plus and 8/10 times I am receiving a long buffer time followed by an error code. I have to ask the bot twice before getting any type of answer. Can someone help me?

[timber shadow]

Custom GPTs can only be used on ChatGPT, but you can make an Assistant "version" of a GPT via the Assistants API and put that on a website. This help article explains the difference a bit more: https://help.openai.com/en/articles/8673914-gpts-vs-assistants

[plush topaz]

Anyone here good at prompting and can help a brother out? haha

[timber shadow]

Hey! #prompt-engineering is a good place for that I'd say. Over there, maybe try sharing:

• What you currently are doing
• What you want the result to be
• What the current result actually is
  This would be a good starting point to get you pointed in the right direction! Lots of helpful folks in that channel

[devout gull]

Hi, is it possible to upload a pdf to chat API? Or do I have to retrieve text from pdf and send it as a prompt

[timber shadow]

Hey, do you mean pass files to the Assistants API? If so, yes, it is possible: https://platform.openai.com/docs/assistants/tools/

[devout gull]

How would it be later used with chat? Or should I use assistant for prompting?

[timber shadow]

Are you using ChatGPT or platform.openai.com, maybe the Playground?

[unique nova]

I'm having an issue with my custom GPTs not being able to access any of the documents in their knowledge base, as well as GPT-4 not being able to read new documents when I upload them. Has been happening for over 24 hours, had never encountered this issue before yesterday.

[devout gull]

I am using open ai chat api

[timber shadow]

I have limited experience with this, so don't want to misadvise you. Maybe try over in #dev-chat if you don't hear back here!

[exotic crypt]

I just came here looking into this same problem

[tacit crow]

Hello,

I’m encountering a persistent problem with my GPT Builder Profile. Despite successfully verifying my website’s TXT record, the system continues to say that the domain is not verified and prompts me to re-verify. This creates a frustrating loop.

Here’s what I’ve tried:

Verification: I’ve verified the TXT record, and the system acknowledges the verification.
Refreshing: I’ve refreshed the Builder Profile page, but the problem persists.
Could you please help me resolve this issue? Are there any known workarounds or additional troubleshooting steps I can take?

Thank you for your assistance.

[unique nova]

seen a couple posts on reddit suggesting we're not the only two

[lapis jolt]

When creating GPTs and uploading a txt document, it won’t read the document. After tricking it into scanning it properly and finishing up the configuration, I get this message over and over:

the recipient must be in the form of tool_name.function_name

[void mist]

Can someone help me estimate prices of OpenAI Assistants API please

[arctic holly]

Is ChatGPT Memories coming to all plus members? I cancelled my membership since at this point there is no purpose in Plus if free members are selected to get features first

[twilit saffron]

ChatGPT is very broken, repeating the same sentences, and when I ask it to guide me through certain things, it does the same thing, for example, "First, do this" and then again "First, do this" and it doesn't stop, and then when I tell it "Done!" it continues to guide me through the first step, what should I do?

[timber shadow]

The memory feature is still in early testing. A week ago they shared that they're just starting with testing with a small group, so most people don't have it yet. The next step we can expect is an announcement for plans of wider release: https://help.openai.com/en/articles/6825453-chatgpt-release-notes

As for free users getting features: very few features of Plus have been added to the free plan. I think only voice and custom instructions? And now seems like probably memory, if initial testing goes as expected perhaps?

[arctic holly]

Right but why are they getting features before Plus members even for testing? I don’t mind features trickle out to the but it’s tone deaf when you have paying members seeing that people who pay nothing are getting features before people who are paying.

[timber shadow]

My guess would be because they plan to release this platform-wide, similar to how custom instructions exist platform-wide. From what I know of memory, it's like a revamped version of custom instructions anyhow, so to me it makes sense. It's not a feature on the same "tier", in my mind, as having access to GPT-4, DALL·E, data analysis, and GPTs.

[vivid mortar]

The second one :) can't do file uploads via the api

[quiet steeple]

Hi Everyone!

My wife did a full asignment with a GPT she built for herself, not understanding, that the preview window conversation would not be saved. She had a lot of work saved in that conversation and hit the back button and lost everything. Can someone help me understand if OpenAI has that convo saved somewhere we might be able to get back?

The assignment is due tomorrow - she is devestadet!

[dark moth]

Does anyone know why one of the circles in my support ticket keeps changing? One turned into an "I", some days later into an "O" and now it's this yellowish color. The ticket is still marked as "not seen yet"...

[thick ivy]

I have been using ChatGPT for research purposes, and it has supplied some excellent answers. However, I've noticed that when I navigate away from certain conversations and return, the answers provided by ChatGPT seem to disappear. This issue started occurring just last week, and it appears that some of the context and answers from ChatGPT are not persisting in the conversations. As a result, I find myself having to redo everything from scratch.

[jagged osprey]

what about the update of chatgpt memory? how can I get it? I have chatgpt4 but the memory feature is only available in another account of mine

[rapid path]

Anyone else getting this Lewis Carolinian nonsense?

[meager locust]

yes, what is going on lmao?

[rapid path]

Someone hacked it perhaps---to poetic and humorous effect.

[meager locust]

This is so insanr

[jagged osprey]

so, useless hacking
bug bounty pay good to avoid this

[daring nebula]

Hi everyone, I was wondering if anyone has come up against an issue where finish_reason = content_filter but not moderation reasons are flagged and te full context window hasn't been hit. Would really appreciate the help with this one.

[rugged wing]

Has anyone had a shared link disabled by moderation when nothing was even flagged?

[potent arrow]

Hi guys! I hope this is simple, but I don't think it will be. Therefore, thanks ahead of time...💯

I'm looking to automate my job to the fullest. My biggest time-consuming task would be compiling release notes. The client wants them submitted on an Excel spreadsheet (one tab for defects and one for enhancements) and PP slides (with each slide composed of tables for each component or tallying data). The thing is that all notes require special formatting and can't have any acronyms. The moment you perform a 'Find + Replace' in Excel, it erases your formatting for any updated cell.

That's why I find it so useful to type the notes in Word, make use of the auto text feature, and apply the necessary formatting (it's a lot).

How can I type and format this information once in Word doc and have that information show up in an Excel spreadsheet and a PP slide - all while retaining all formatting?

[past geyser]

I'm not sure this relates to OpenAI! At least I'm not aware of any GPT implementations that can help you do that.

It can potentially reformat text for you but not with that level of integration with Microsoft software afaik

[winter ingot]

hi guys, my chatGPT is running verrrrrrrry slowly, almost unberable, I am from Malaysia

[west tusk]

Hi all. I want to report an issue I've encountered with the ChatGPT service on my desktop, which happened after subscribing to GPT-4.

Problem Description:
After entering text into the ChatGPT interface and pressing Enter, there is no response, and the input text simply vanishes. This issue prevents me from using the service entirely.

Troubleshooting Steps Taken:
Refreshed the webpage
Cleared browser cache and cookies
Disabled browser extensions
Tried using different browsers (Chrome, Microsoft Edge, Safari)
Checked for browser updates and ensured the latest version
Attempted to access the service on a different device (It works on mobile app)

Device and Browser Details:
Device: MacBook Air 2020
Browsers Used: Chrome, Microsoft Edge, Safari
Internet Connection Type: Wi-Fi
Subscription Type: ChatGPT Plus

I have followed all recommended troubleshooting steps, but the issue remains unresolved. I would greatly appreciate your assistance in identifying and resolving this problem, as I rely on ChatGPT for various tasks.

Has any of you ever faced this issue?

[last belfry]

Hi everyone! Is there a place where we can ask help or support from OpenAI ? Or is it just here, where we ask questions and hope that someone from OpenAI is listening ? Their website says there is a "help" button but I searched everywhere and there is nothing. And I'm paying a subscription too, which is very anoying.

[sweet ravine]

You can if you login in your account and click on get help

[timber shadow]

Hello! Since it's working on another device, let's troubleshoot the device with the issue. Do you have a VPN installed on the MacBook?

[timber shadow]

Hello, the chat widget on help.openai.com is the official support channel. Open the chat widget in the bottom-right of that page, click through the most relevant options, then you'll eventually find a path that lets you type your own message. This will submit a ticket that OpenAI support can review and get back to you on. It may take some time, possibly days. You'll get an email when they reply to you!

[west tusk]

I am using the same VPN on phone and laptop
Because it will not load in my region without VPN :/

[rugged wing]

I have a very simple question:

How can I Remove My OpenAI Account?

[timber shadow]

Are you located in one of the countries listed on this page? https://help.openai.com/en/articles/7947663-chatgpt-supported-countries
If so, you should be able to access no problem, so we could try some other troubleshooting if you'd like. If not, that might be the source of the issue.

[wooden fjord]

Hey community people 👋

I have noticed that the OpenAI Chat API returns the following for "what X is it today?"

// [current time]
// [current day]
// [current year]

It seems like it is deliberate for possible localisation to occur on responses.

Anyone know what these are referred to, and if there is a documented list for post processing responses? Are they embeddings? or something else?

[timber shadow]

Hello, this article lists the steps for deleting your account! https://help.openai.com/en/articles/6378407-how-to-delete-your-account

[west tusk]

I see. thanks for your help

[rugged wing]

Thank you very much for your information @timber shadow!
I appreciate it!

[trail cairn]

thats probably just what it learned it during training and it will output that if you dont give it the current time

[wooden fjord]

if you ask in ChatGPT it gives you the date. Does that mean, despite telling it to ignore previous instructions, it is still getting some context?

[trail cairn]

date in the system role message is more powerful than user message so it wont completely ignore whats in there

[reef zodiac]

Hi, I am trying to build a resume parser using GPT4 turbo model API but I am unable to figure out which file API to use, I basically want to provide it a pdf file and it should parse and return me text data.

This is the documentation I am using
https://platform.openai.com/docs/api-reference/

[wooden fjord]

nice, and thanks 💪

[tight prism]

I haven't been able to use gpt-4 via the API for the past day or so, via the playground or otherwise. I was able to do so up until a couple days ago. I have a paid account, etc. gpt-4-turbo-preview and the other models work fine. any suggestions?

on the playground, I get the error "The server had an error processing your request. Sorry about that! You can retry your request, or contact us through our help center at help.openai.com if you keep seeing this error."

[wooden fjord]

oh it told me 😂

[soft fern]

Hi everyone, anyone else unable to use Android app? Another user reported this problem as well. It just loads infinitely on open.

[fluid whale]

Anyone else so overwhelmed with GPT-4 that it completes the task correctly at first but then just spouts nonsense from the middle onwards?

[full surge]

did you notice the chatgpt4 not accurate like before or mine have problems?

[full surge]

mine but spouts nonsense from the start

based on the current chatgpt4 result i think it is a good time to fire the Sam Altman

[ocean lake]

I need help getting help! ChatGPT deleted ALL of my Archived files. I've left a message on as a bug report in that forum, but no response. And since it doesn't exactly meet their criteria for posting bugs, I fear they are just ignoring it. How do I get my archive files back? This was their mistake not mine! I know they store chats for 30 days after deleting so can't they restore my archived files? Somebody please respond point me in the right direction! Thanks

[timber shadow]

Hello! Sorry this happened. You might try reaching out to OpenAI support at help.openai.com. Open the chat widget in the bottom-right, click through the most relevant options, then you'll eventually find a path that lets you type your own message. This will submit a ticket that OpenAI support can review and get back to you on. It may take some time. You'll get an email when they reply to you!

[ocean lake]

thank you I appreciate your response. I did what you have suggested, and yes I have not heard back from them. I'm sure they're busy and understaffed for this sort of thing. But you can imagine I had some pretty important chats in there that I had not locally backed up yet. My bad!

[unkempt nexus]

I'm getting a "Gateway Timeout Error" when saving my first GPT after just signing up for Pro. Is this a known issue?

I think I can use the draft/preview view for now but it'd be nice to know that all the work I just put into customizing the GPT doesn't go to waste. This might be because I have large specification documents uploaded, but not sure. Would be nice to know if the documents are too large.

Aaaaand just like that, after trying for an hour, it finally worked lol

[astral wren]

I need help with sora

any one can help out

[rigid musk]

?

[astral wren]

i am trying to use it but every time i log in it sends me to chat gpt

[cold folio]

Hello! Sora has not been released yet.

[astral wren]

theres a way to use it, thats why i even tried in first place

[rigid musk]

sora isnt officially out

[astral wren]

ouch so i guess then buddy got the hack for it or early use?

[cold folio]

It's not released yet. please only trust on informations from OpenAI website and Discord
https://openai.com/sora

[spark rune]

Whenever I use openAI API I get a message saying 'You exceeded your current quota' when I havent spent any money for the last 3 months and I check and i havent reached my limit

help🙏🙏🙏

[rough imp]

Do you guys know how to link the chatbot from the assistants module to a webpage through VSCode?

[empty vault]

people! Chatgpt has gone haywire! I have a deadline and it's messing with me, intentionally for sure. I was like ok...may be server problem, and trying have my calm but no, not one day, second day, third day, even fourth day, close to second week now....it did Elastic regression once for one dataset, then i said do replicate this method for the another dataset and it just kept saying "it can't be done here in this environment, kindly do in your software python or R" "I can show how it can be done"

can anyone tell me where the hell is openai's servers so that i can go and crush them into pieces!!! Their AI is driving me nuts!!!

[frigid gull]

when i ask chatgpt to rewrite my code or fix something etc.. i go to alt tap on something else until its finish but i see a button "Continue Generating" can i make it just generate and continue the code without confrim?

[frigid gull]

please mention me when you answer ebcause i dont get notice

[boreal pecan]

it seems like chatGPT will not show "doing research from website" prompt when he doing this now?

[thorny belfry]

Hey everyone,

I think I've figured out why I've been having troubles with ChatGPT 4. So, I'm on a corporate VPN when I access work-related stuff, including my ChatGPT Team account. Here's the thing: whenever I'm connected through our default VPN server in Germany (Hetzner DC), I run into issues with ChatGPT4 (oddly, 3.5 doesn't give me any trouble). But if I switch my VPN profile to France, voilà, ChatGPT4 starts working perfectly.

Is anyone else experiencing this? Could someone validate my hunch? Maybe it's a common issue?

I've tried reaching out to ChatGPT support, but haven't got much help, it's very unresponsive. Bot just suggested turning off the VPN, which isn't an option for me since it's a company policy and my ChatGPT Team account is considered as a corporate.

[languid stratus]

I have the same problem. It worked for months but today it stopped working 😦

[spark rune]

you gotta buy credits

[spring copper]

hello every one i want to train chatgpt with html and css files . is that possible ?

[sage geyser]

there is a tampermonkey script to auto click this just look it up on github adamlui/chatgpt-auto-continue

[rustic marlin]

yep, vpns are known to cause issues

[past bloom]

hello I have a project I want to use opena Ai to generate html from text and the page should use components that I provided in the prompt
the problem is the style of the page is not good enught but when I change the prompt to make a nice ui design for the page open ai generate a code and ignore the components that I've sent

what is the best solution in this situation

[mossy mesa]

You should probably take this up with your IT, who can likely set up a specific policy in the company's domain for ChatGPT. And if not, they might need to re-think their Teams ChatGPT strategy since it and VPNs don't always get along.

I want to report a bug with the chatgpt beta memory feature and I've been having some difficulties with getting posts to stick on the topic, since I can't reproduce it. How should I go about making my good faith report, since my original got deleted?

[cunning orbit]

If I try to generate a word document. ChatGPT only can generate 1-2 pages and then it stops always. Any tips on how to change that? I alrdy said to do it step by step for every page

[mossy mesa]

It has a response budget. The solution is another prompt. Usually "Continue." is enough.

[timber shadow]

Hey Darth, maybe this would work:

Title: Had memory feature, now missing

Steps to reproduce:

1. Got the memory feature
2. Tried to add data in an XML file to memory
3. Shortly after this, memory feature disappeared

Expected result:

Memory feature to remain

Actual result:

Memory feature disappeared

Additional information

Browser: [your browser]
OS: [your OS]

[mossy mesa]

Thanks. I'm genuinely happy to follow guidelines.

I like the steps to reproduce. I mean they're impossible, but I like them. Thanks again.

[timber shadow]

Sure thing! If you think it might help and if you want to, you might share the chat link for the chat you used to try to import the XML file, too. In the template, that would go above "steps to reproduce"
#1202282235877916683 message

[frigid gull]

oh really! thank you for this.

[mossy mesa]

The XML contains the dev journal for my entire collection of custom GPTs. It's only about 8 or 9kB, because of sparse priming representation summarization, but it's not for sharing, regrettably.

While the product is open source, my dev journal isn't.

Since the chat responds to the context, well, I can't provide the link to the relevant chat.

But it's about 8kB or maybe 8.5kB of properly formatted sparsely tagged XML. The most advanced data are unified numerical time/date stamps.

[timber shadow]

Makes sense! Alternately, you could share just the chat ID from the URL without making it a shared chat. An OpenAI team member would still be able to look at it, without you making your work public. The chat ID is the string after "/c/" in the URL for the chat. Example: 220d0d79-8283-4806-88a8-d51c679a5924

[mossy mesa]

I may go back and look at it if I get engagement. With my vision problems, you can imagine the struggle that'll be. :\

corneal keratitis sucks.

[timber shadow]

Also just generally, I do still think it's possible that the feature appearing and disappearing so quickly for you was just an unfortunate byproduct of simple temporary A/B testing on the part of OpenAI, but regardless, I hope you and all of us get to try the feature soon!

[mossy mesa]

That's possible. I think framining is important, though, too. Especially when I had a number of correlating questions that were nontrivial. I hope you understand how devastating it is to see those questions removed.

I left them out of the new report, but they remain.

[slate dawn]

hello is chatgpt down ?

[timber shadow]

Hello! It's loading normally for me and there's no message on https://status.openai.com, what kind of issue are you having?

[timber shadow]

Can you try in a private window with all extensions disabled?

[slate dawn]

it's good i just changed my network and it's working now thank you

[hard knot]

Do you know if I can mix yearly and monthly plan for chat gpt teams? For example buy 10 seats on a yearly contract but add/remove additional users on a flexible/monthly plan as I need it?

[regal lynx]

I don't know all of the ins and outs of the team plan. I'd be interested to see if it supports yearly and monthly plans simultaneously

[timber shadow]

This is a helpful article on this topic!
https://help.openai.com/en/articles/8792536-manage-billing-on-the-chatgpt-team-subscription-plan
Excerpt:

Example: A user buys 5 seats at an annual price of $25 per seat, totaling $1,500. On day 10, they add 3 seats, now totaling 8 seats. On the next invoice, they are charged $875 total, covering the cost of the additional seats for the days used and the remainder of the year.
Invoice 1: 5 seats * $25 * 12 months = $1500
1st Monthly true-up cycle: 20 days used /30 days * 3 seats above baseline * $25 monthly subscription + 3 seats above baseline * $25 monthly subscription price * 11 months remaining in annual plan
The owner removes 2 users from a workspace on the 20th day of Month 2. Nothing changes in the billing process since it is below the baseline.

[crystal shore]

Its been like 3 days now when I go to it its like that.

Id like to remove my payment details

It works just fine in my phone though.

Ive been a power user of the gremore maybe thats why I got bloked?

[timber shadow]

Hi, if you're just trying to remove your payment details, can you do it from your phone at https://chat.openai.com/#pricing? Then "Manage my plan" in that modal.

As for why it's not loading on the computer: you're almost definitely not blocked. To troubleshoot that, can you first test in a private window with all browser extensions disabled?

[crystal shore]

I have no extensions in my latest chrome, tried it in incognito as well no luck. Also in my phone I cannot access my subscrition itl say that I have used a different platform when I paid something like that.

[timber shadow]

On your phone, can you try it in the browser instead of in the app?

[crystal shore]

It asks if I was a human by cloudflare 2 times.

The same

[timber shadow]

All right, I'm wondering if it doesn't like something about your WiFi network then. Can you try:

• Disabling WiFi and connecting to your cellular network
• Opening a private window of your browser
• Signing in fresh?
  Could help identify whether it's a network thing, and a private window will help avoid any potential cache issues.

[crystal shore]

What the heal, it works. So its my external IP?

[timber shadow]

Good to know! It might be some setting of your WiFi network then, yes. You might check your DNS settings as that's a common culprit, or maybe even just try a simple off & on again cycle.

[crystal shore]

Ok thansk

[drowsy wedge]

is it normal for streaming requests to be faster in response time while compared to non-streaming requests?

streaming requests takes ~800 milliseconds to generate a response before start streaming it, while non-streaming requests take 7-8 seconds to generate a response (same prompt)

oh, forgot to say, this was using the API

@tribal talon could you possibly help me? just saw that you allowed mentioning you

[tribal talon]

Sorry I’m not at home right now to look at stuff

[drowsy wedge]

okay

[deep hare]

Hi, was translating some technical documents by ChatGPT-4 and suddenly I can’t send to ChatGPT, there is no response when hitting send. Is my IP blocked cause it works if I change location to my work-PC with another public IP.
At home I have tried on several computers but same result, also different web browsers.

[timber shadow]

Hi, do you have a VPN running normally? Just wondering about why it's working on one network and not on another.

[tribal talon]

That does seem normal to me. That’s kind of the point of streaming it.

[deep hare]

different ISP

[drowsy wedge]

so it generates it while streaming?

and also, I was getting way lower response times before

tried different IPs too, same thing

[deep hare]

work with one ISP on several computers, but where i was translating all computers cant get response from chatGPT,

[tribal talon]

I believe to an extent it is yes. Not sure on exact technical details but normally it sends back entire response at once vs streaming it sends it as it generates. Piece by piece. So it will get you info “faster” in a way

[deep hare]

can log inn but all i type just dont send

[drowsy wedge]

is there any way I can make the non-streaming response times faster? I'm using gpt-4-0125-preview

it was faster before

[tribal talon]

No idea. That sounds like out of your control

[drowsy wedge]

yeah, probably

[timber shadow]

Can you right-click the page and click "Inspect", then switch over to the Network tab? Does anything stick out there as a clue as to what may be blocking use of ChatGPT? Probably don't share that screen here as it could have potentially identifying info in it, but it might be helpful for your own troubleshooting.

If it's working on one network and not on another, you might check the DNS settings of the problematic network too. That is a common culprit of issues like this.

[deep hare]

DNS setting has not changed, so its not DNS issue, will do the inspect and see, but im realy sure its some kind og IP block.

[indigo fulcrum]

to work with gpt is like to work with a worker that always forgot the task all the time

[grand karma]

how do you get chatgpt plus? when i try to buy it the page just shows my payment method, billing information and invoice history. there is no button to purchase

[supple atlas]

I'm having issues performing the node js api call. I set the API key in the environment variables. I verified that i’m on v4.28 of SDK. The error suggests that the chat property is being called on an undefined object within thesendToGPT function. Any input/help would be much appreciated

[timber shadow]

Hi, can you share a screenshot of what you're seeing on this page? Be sure to edit out any personal information in the pic first.

[timber shadow]

Hi! If you don't hear back here, #dev-chat might be a good place to ask too!

[timber shadow]

Is there anything above "Payment Method" that says "Current Plan"? When you were previously subscribed, did you subscribe via the web or via the iOS/Android app?

[grand karma]

no there is nothing at the top nor the bottom,
im not subscribed because i have issue with card being declined before

[timber shadow]

Do you know what caused the entries under "Invoice History"? Just wondering why there would be anything there at all if your card has been declined.

[grand karma]

i was subscribed last year twice

but recently my card keeps getting declined, ive reached out to support they told me to try again
but as you can see the ui is completely different than before

[timber shadow]

Hmm, odd! Can you test in a private window with all browser extensions disabled? Or perhaps in the browser on a phone?

[grand karma]

the same thing shows up

oh it seems to work now

idk what changed but it didnt ask me to do the audio test this time and pops up the old payment screen

thanks!

[tacit vault]

Hi guys,

i have some issues regarding the tokens-calculation in my assistant_api and i need some help urgently!

I have created an assistant based on the GPT-4-1106-preview model with custom knowledge base (currently two pdf files).
Until now i did not have any issues, but since last month, the tokens-calculation regarding my input (question) got so drasticly high that it cost me way too much money.

My normal use case is that i enter an input of max. 150 words and get a response of max. 500 words.
Until now, whenever i asked a question with one sentence (4-5 words) and got the response of like 350-400 words the amount of used tokens was round about 700-900 tokens, which is perfectly fine!

Now if i ask the same question with the same 4-5 words, it counts the input_prompt lenght up to over 30.000 tokens !!! But the output tokens are still calculated accordingly !
So the main problem is that the input_prompt somehow adds thousends of tokens to the actuall question_prompt!

Can someone please help me fix this or is this a bug that needs to be reported to openAI ?!

Thanks in advance!

[flint sail]

Hey I also have a Question,

I'm developing an App in Android Studio that will first use OCR to extract text via Google ML Kit. The text will be stored into a JSON file and then passed into GPT to try and categorize what is on the image. I want to condense this step down into just sending an image, could anyone tell me if this is possible to send an image while using Android Studio? Will I have to encode and send the image? Will it cost a lot of tokens? Is it gated by a certain AI model I dont have access to? Image attached is what I have so far using http3 on android

[willow canopy]

I’m dealing with this issue. Keeps saying unable to load site and unblock cloud flare challenges. Does anyone have an idea how to fix it?

[bitter birch]

For two days I'm getting this error! Tried up to 6 times. What is happening?

[vast junco]

You can take a look at this article, https://help.openai.com/en/articles/8184038-captchas-in-chatgpt

While encountering CAPTCHAs occasionally is normal, you shouldn't constantly encounter them. If you do, here are some initial troubleshooting steps to help resolve the issue:

• VPN and Browser Configuration: Make sure you're not connecting through a VPN and check for any non-standard browser configurations, such as browser extensions that might interfere with the page.

• Incognito Mode: If you've confirmed you're not on a VPN, try using an incognito window to see if that helps. Unusual behaviors can cause CAPTCHAs to show up more frequently, so using an incognito window can be a good first step to address this.

[vast junco]

you can contact the support by following this link: https://help.openai.com/
there should be a chat button at the bottom right. you can explain the issue and you should get some support quickly

[vast junco]

when there is a custom knowledge base (documents, files, PDFs), the model can use retrieval to read the files. Those files effectivement counts as input tokens as part of the context. You can read more about the process here: https://help.openai.com/en/articles/8868588-retrieval-augmented-generation-rag-and-semantic-search-for-gpts

[vast junco]

OpenAI offers the GPT-4 Vision API which can receive images and analyze them. From my experience, the OCR capabilities of GPT-4V are not always reliable, but it works ok with simple english text. I haven't tested all parameters present in the API though, I know we can post a higher resolution (you can refer to the pricing to avoid surprises.)

[rugged wing]

with which AI are the graphics in the dell-e-gallery channel created? https://discord.com/channels/974519864045756446/1154829862171844679

[tardy ether]

?? anyone knows what happen?

[sand solar]

anyone knows how to build a openai chat bot in discord server with no-code required?

[rigid musk]

Also #off-topic

[timber shadow]

Hey! They're made with DALL·E, mostly DALL·E 3. You can try DALL·E 3 with ChatGPT Plus, with Bing's image creator, or in the #image-bot channel with five free images a day!

[timber shadow]

Hi, can you try in a private window with all extensions disabled as a test?

[gritty trail]

How to fix this?

[trail cairn]

try in private mode

[spark rune]

where can I use gpt4 0314?

[trail cairn]

only via api and it will be gone in few months

[spark rune]

send me links

of website where i can use it

or where the api is

[timber shadow]

Hi, Radium is the expert here so they may know more, but I did find this:
https://platform.openai.com/docs/deprecations/2023-06-13-updated-chat-models

As of 01/10/2024, only existing users of this model will be able to continue using this model.
So if you haven't used the model before, you may not have the option currently.

[trail cairn]

you need to buy credits to use it here https://platform.openai.com/playground?mode=chat

[trail cairn]

ah nice i have missed that 🙂

[loud flax]

Heya, do you guys know of any good method to save a copy of a conversation? (Other than copy-pasta via select & ctrl c)? I'd like to retain some conversations for reference afterwards but the UI is frankly terrible for this sort of thing.

[ebon ledge]

ChatGPT saying my message is too long to be submitted.. wth what’s going on it’s usually taken much more before ??

[quiet dew]

Tokenizers can split up tokens in surprising ways. Sometimes whitespace is 1 token, sometimes it is many.

[hazy sand]

I don't know if anyone is having same problem here but my chatgpt plus stopped generating images. is it just me right now?

[charred parcel]

just tried it and it's working for me

[hazy sand]

surprisingly, now it's working for me lol

[waxen kiln]

Hi, I need to know what I'm doing wrong to get "Our systems have detected unusual activity from your system. Please try again later." I haven't changed my usage pattern or added any new chats. I can use my workspace ChatGPT but my personal one keeps giving me that message on old chats or when I start a new chat

[paper maple]

I'm getting the same error. I want to speak to an actual person and not chatGPT. This is annoying.
I'm getting Our systems have detected unusual activity from your system. Please try again later.

[white mist]

Does anyone know how to solve this in the assitant API?

[waxen kiln]

I would just like to know if it's something on my end or theirs, I don't use VPN and my office is SOHO, no proxy, no other routers to go through. I haven't changed since I started using it.

[paper maple]

yes. Same here. I've not changed anything either. Its been working just until a few hours ago.

[jade pagoda]

can anyone tell me how to get in touch with Nitro support? I paid for a subscription but when I try to use it, it keeps telling me to join.

[covert fiber]

Hi, entered here and discovered many of us getting unusual activity errors

if I switch to company's account inside same login, I can use it

[waxen kiln]

thank goodness. I thought i screwed something up. I wonder if someone who only has a personal account and no workspace account has the same issue

[flint sail]

But you can pass an image through? Like from my example below im just sending text / json after using OCR to extract it. I would rather ust GPT-4 to pass the image to eliminate that step because the OCR I am using is not accurate. When passing the image manually myself into GPT-4 it provides me a perfect categorization of the image I am sending

[covert fiber]

My account with personal email works. With my company's account works when I use it in non-personal subaccount mode. All of this on same browser. And issue happens also with incognito mode. I have no active VPN and never asked for illegal stuff.

[worldly geyser]

is there any way to surpass the message cap limit by logging in on a different device or something

i don’t want to wait 3 hours to use the app again

[vast junco]

You will need to encode your image in base64 if you want to upload it directly to GPT-4V. You can refer to this guide: https://platform.openai.com/docs/guides/vision

[timber shadow]

There is not, the usage cap is per account. #openai-chatter message

[brazen thistle]

Anyone got issue with uploading pdf documents right now?
You guys may have a try, IDK why GPT4 is keep failing reading PDF documents at this moment

[flint sail]

Thank you thank you. I got it to work. I'm so sad I decided to work with Android Studio and not just do a progressive web app ._.

[vast junco]

hm, it's relatively easy to convert from one to the other, if your project is still small. ChatGPT can probably translate the code from one language to another. You will probably want to look at frameworks like AngularJS, ReactJS, VueJS, or anything similar which can make web development easier.

(edit: for multi-platform application, ElectronJS is a popular framework. Works with Angular, which is a well known and mature framework)

If you need more help with development, feel free to visit the #dev-chat channel

Even Discord is built with Electron. I think it's using the ReactJS framework with Electron

[brazen thistle]

It appears that the file you've uploaded, "XXXXXXX.pdf," is either empty or in a format that I'm unable to read. Could you please check the file and try uploading it again?