[Bug]: All my accounts deactivated
sromexs opened this issue · 102 comments
Is there an existing issue for this?
- I have searched the existing issues and checked the recent builds/commits
What happened?
ALL of my accounts that used in this api are deactivated 🤦♂️
Who has this problem too ?
You should log out and login again to see if your account deactivated or not
Steps to reproduce the problem
- Go to ....
- Press ....
- ...
What should have happened?
...
Version where the problem happens
...
What Python version are you running this with?
3
What is your operating system ?
No response
Command Line Arguments
...Console logs
There is no cmdAdditional information
No response
+1
Same
Error in ask: {"detail":{"message":"This key is associated with a deactivated account. If you feel this is an error, contact us through our help center at help.openai.com.","type":"invalid_request_error","param":null,"code":"account_deactivated"}}
+1
+1
+1
+1
+1
OpenAIAuth.Error
It's happening to a lot of people. Seems they can detect anomalous usage
And please don't spam the +1 button. Use thumbs up reaction to the initial issue
It's happening to a lot of people. Seems they can detect anomalous usage
Maybe the bypass method was detected by
OpenAI
If they detected it, they could just stop the method rather than banning accounts that use it. None of my accounts have gotten banned for some reason
I don't think it's because of the problem of bypassing Cloudflare. I have received feedback from users that using GPT 3.5 Turbo was also blocked. And this is the official release of the API.
我的老号被封禁了，好伤心，有没有什么办法解封？
I don't think it's because of the problem of bypassing Cloudflare. I have received feedback from users that using GPT 3.5 Turbo was also blocked. And this is the official release of the API.
People getting banned using GPT 3.5 is due to multiple accounts. This library uses the official API for V3 and makes no difference
The main issue is with V1 right now. I'm testing with my accounts to see if I can determine the trigger for ban
Haven't used the library for around 3-4 days and got banned today. Not sure if this info helps.
Haven't used the library for around 3-4 days and got banned today. Not sure if this info helps.
The current bypass has only been around for 3-4 days. I think they must be tracking usage and cracking down on anomalous accounts.
I have a dozen accounts for testing. I'll let everyone know when I get banned. For now, use V3 to avoid it
我有十几个帐户要测试。当我被禁止时，我会让所有人知道。现在，使用 V3 来避免它
I have a lot of accounts and I can provide some to you if you need them. Help you test.
Thanks. I have enough for now. I'm testing with different factors - Different JA3 fingerprints, source IP address, accounts, etc
People getting banned using GPT 3.5 is due to multiple accounts. This library uses the official API for V3 and makes no difference
I've only used the API occasionally (less than 20 requests a day) and I've only used a single account, yet my OpenAI account just got banned today. Whatever bypass that's being used is definitely being detected and logged by OpenAI.
People getting banned using GPT 3.5 is due to multiple accounts. This library uses the official API for V3 and makes no difference
I've only used the API occasionally (less than 20 requests a day) and I've only used a single account, yet my OpenAI account just got banned today. Whatever bypass that's being used is definitely being detected and logged by OpenAI.
This is for V1. We know they're banning V1 users. I mean V3 isn't getting banned
face the same issue, my old account was deactive today, And i never use apiKey or accessToken for opensource project before, but yestoday I use the proxy address, my account deactive this morning.
This is for V1. We know they're banning V1 users. I mean V3 isn't getting banned
Ah, I should have read your message more closely.
+1
@acheong08
My proxy has not been disabled, is it because I didn't log out? I have another account with the same configuration, but now it's reporting OpenAIAuth.Error. This is the proxy that has not been disabled for me.https://www.postbook.xyz/read/chatGPT
+1
Even my "Login with Google" account deactivated.
confirmed, my account deactivated too
So do you have banned accounts with Plus subscription?
+1
me too
Even using V3 is also deactivated. Maybe it is because of the account has telephone in country/region A (eg. U.S.), but your IP address is in country/region B (eg. HK). Just suspicion. Another possible reason is that only IP address from U.S. is allowed, try using a U.S. proxy.
Even using V3 is also deactivated. Maybe it is because of the account has telephone in country/region A (eg. U.S.), but your IP address is in country/region B (eg. HK). Just suspicion. Another possible reason is that only IP address from U.S. is allowed, try using a U.S. proxy.
My account that was made and used in the same IP got banned.
即使使用 V3 也会停用。可能是因为该帐户在国家/地区A有电话（例如。美国），但您的 IP 地址位于国家/地区 B（例如香港）。只是怀疑。另一个可能的原因是只允许来自美国的IP地址，请尝试使用美国代理。
那样的话岂不是一大堆账号会被停用？
即使使用 V3 也会停用。可能是因为该帐户在国家/地区A有电话（例如。美国），但您的 IP 地址位于国家/地区 B（例如香港）。只是怀疑。另一个可能的原因是只允许来自美国的IP地址，请尝试使用美国代理。
那样的话岂不是一大堆账号会被停用？
My server is deployed in Singapore, and the same problem occurs
does anyone know why they were banned?
My four accounts deactivated today.
I use V1.
same. deactivated 4 accounts sigh. V1.
+1, mine was deactivated too [cry]
does anyone know why they were banned?
perhaps due to an incorrect approach taken by the proxy server, a
tokenandpuidmismatch was detected.
But my 'Google Login' account deactivated also. (it cant be used with code. only browser usage.)
does anyone know why they were banned?
perhaps due to an incorrect approach taken by the proxy server, a
tokenandpuidmismatch was detected.
I've migrated to a new approach and people are still getting banned
I'll try out something else tomorrow. There are a ton of ways to get around Cloudflare.
i have the same problem
Just to understand the scope of this. Is this affecting only accounts that connected via proxy? Or also the ones that did it via access_token exclusively?
Same here. Yesterday I pay for Plus Subscription, I used ChatGPT-4, and all was normal. Today I have seen that my account it's "deactivated". I send an email to support, let's see if they respond (if someone receives a response from support please comment here!)
Just to understand the scope of this. Is this affecting only accounts that connected via proxy? Or also the ones that did it via access_token exclusively?
I think this is access_token?
is that chatGPT is not working or you're getting an error that your account is deactivate?
Everyone please stop using V1 while I sort this out
wait for more info
I only have one openai account, and I used my real US phone number and my google account to register it. My ip address is from the US, but I guess it's irrelevant here because this library uses a non-US proxy? Got deactivated yesterday after trying v1, and my phone number could not be used to register any other new openai account anymore.
I did not get any email from openai stating I am banned.
Here's my situation: trying to log into openai for the first time today. But after only a few minutes of successfully using chatgpt, I was emailed that my account was banned and I couldn't log in.
I am not using any api related functionality. It was the Hong Kong ip when I tried to register for the first time, and then I switched to the US ip after the webpage showed that it could not be used in this area (the mobile phone number used a text message received from Indonesia). It was quickly banned.
There is very little discussion of this situation on the web. I want to know how to solve this problem.
这是我的情况：今天第一次尝试登录openai。但是在成功使用chatgpt几分钟后，我收到了电子邮件，我的帐户被禁止，我无法登录。我没有使用任何与 api 相关的功能。第一次尝试注册时是香港IP，然后在网页显示无法在此区域使用后，我切换到了美国IP（手机号码使用了从印度尼西亚收到的短信）。它很快就被禁止了。网络上很少讨论这种情况。我想知道如何解决这个问题。
正常使用应该不会出事吧，就是在官网使用
这是我的情况：今天第一次尝试登录openai。但是在成功使用chatgpt几分钟后，我收到了电子邮件，我的帐户被禁止，我无法登录。我没有使用任何与 api 相关的功能。第一次尝试注册时是香港IP，然后在网页显示无法在此区域使用后，我切换到了美国IP（手机号码使用了从印度尼西亚收到的短信）。它很快就被禁止了。网络上很少讨论这种情况。我想知道如何解决这个问题。
正常使用应该不会出事吧，就是在官网使用
但我确实如此，参考了网络上的使用ChatGPT注册教程，也没有遇到什么太大的问题。并没有调用什么api和其他的操作，也不是开发者……很奇怪。网上没有这个的解决方案和原因。我找半天才发现这个帖子里的讨论似乎和我的情况比较类似……
If you use a public-sms platform for registration, probably other accounts registered with that phone number got banned, resulting in your ban. It seems that openai uses the phone number as a user-identifier.
You registered to openai from A Country. if you try to login different location(Example B country) OpenAI saying you cannot login from B country. Because you registered from A Country.
So, we registered From A country and we send requests trough proxy server(server location X country). Now OpenAI realized this stuation, we are logged from Our countries and sent request through Proxy server via access_token. We all banned.
Setup your proxy server location in your country(Where you registered accounts.). You will won't be deactivated. (That's my opinion)
You're welcome to try: https://github.com/acheong08/ChatGPT-Proxy-V4
Setup your proxy server location in your country(Where you registered accounts.). You will won't be deactivated. (That's my opinion)
I'm working on a different method that'll hopefully make it safer but it'll take a while (a lot of work)
I'm working on a different method that'll hopefully make it safer but it'll take a while (a lot of work)
I am not blaming you bro. You did your best. I appreciate
Setup your proxy server location in your country(Where you registered accounts.). You will won't be deactivated. (That's my opinion)
perhaps not. i also have a running bypass server, and no one has been banned because of it.
Different method?
Or maybe the usage of mine is too great and became suspicious?
Setup your proxy server location in your country(Where you registered accounts.). You will won't be deactivated. (That's my opinion)
perhaps not. i also have a running bypass server, and no one has been banned because of it.
Different method?
your proxy service lacks some additional headers.
Mind sharing them?
I've checked and don't see any significant other ones.
I'll add everything just in case
by the way, the puid and token in your bypass server do not match, which is very easy to detect.
Yeah I know that. How are you getting around it or are you only using plus accounts?
I'm migrating to a cf_clearance based bypass later on
perhaps not. i also have a running bypass server, and no one has been banned because of it.
Can people use your bypass service while I fix that?
perhaps not. i also have a running bypass server, and no one has been banned because of it.
Can people use your bypass service while I fix that?
i have no traffic control, and it might crash :(
Can you perhaps share it with me privately then? I think I'll need to take down my current server to prevent further bans. If it's efficient, I should have the resources to scale it
I have both a plus account and I been using v1 (and v3 with another bot), no ban as of yet. Very weird you all got deactivated. The account has my personal phone number (no voip) etc.
We actually don't know the exact reason for the ban right now. I have now 12 accounts I use interchangeably with the API including one plus account and only one has been banned so far, that one being my oldest account. The exact reason for the ban isn't clear to myself either as I have broken a lot of their policies previously with the "jailbreak" prompts
We actually don't know the exact reason for the ban right now. I have now 12 accounts I use interchangeably with the API including one plus account and only one has been banned so far, that one being my oldest account. The exact reason for the ban isn't clear to myself either as I have broken a lot of their policies previously with the "jailbreak" prompts
Well I deleted my oldest account and I have been using jailbreak prompts regularly. Not sure when will ban come if it does come. But it makes sense that we do not know the exact reason.
is there any hope to appeal for re-activating the account?
Dear user,
According to the privacy protection laws in the United States, such as the California Consumer Privacy Act (CCPA) and the General Data Protection Regulation (GDPR), users have the right to access, correct, and delete their personal information, as well as request service providers to export their personal information. However, whether information service providers need to provide user information export services is not a legal requirement, but depends on the policies of the service provider. Some service providers may offer this feature, while others may not.
ChatGPT
my account was deacitved,Does a complaint work?
我的帐户被取怎么投诉？
@acheong08 我的代理没有被禁用，是因为我没有登出吗？我有另一个具有相同配置的帐户，但现在它报告 OpenAIAuth.Error。这是没有为我禁用的代理。https://www.postbook.xyz/read/chatGPT
I also encountered the same problem, OpenAIAuth.Error, did you solve it?
is there any hope to appeal for re-activating the account?
According to US privacy protection laws, such as the California Consumer Privacy Act (CCPA) and the General Data Protection Regulation (GDPR), we as users have the right to access, correct, delete our personal prompts, and request OpenAI to export our personal prompts.
ChatGPT:
As a language model, I do not handle or store user's personal information. OpenAI strictly complies with applicable privacy protection laws and regulations, and makes every effort to ensure the security and privacy of user's personal information. If you have any requests or concerns about your personal information, you can contact OpenAI's privacy team. They will respond to you as soon as possible and provide support.
my account was deacitved,Does a complaint work?
Nope. They will ignore you just like all other corporations / companies do when you get banned.
When will there be an update on this matter announced? @acheong08
If an account is disabled due to this, the entire IP address is banned. When I tried to create a new account from the same PC, the account was deactivated the moment I logged in
Correction, I tried it via VPN and it was the same.
I posted on reddit about my ban.
https://www.reddit.com/r/OpenAI/comments/11u6t1o/openai_deactivated_my_account/
I recommend that you do the same, in reddit or any social media, as it is othe only thing that will make OpenAI more aware on how they treat users and customers.
When will there be an update on this matter announced? @acheong08
I've swapped my back end a few days back to a new approach but it seems more people are still getting banned. I can't to figure out what the exact cause of ban is. With around 12 accounts, I've tested a few factors:
Bypass method: cf_clearance, puid, another unprotected endpoint
Frequency of messages: Set at 1/hour, 10/hour, 20/hour
IP source address: Azure, Oracle, DigitalOcean
Account bans: 0 from the testing group. 1 personal account banned.
Note: Each of these used the bypasses locally.
Personal account used public proxy server.
Possible conclusion: Access from the public proxy servers' IPs were logged and blocked. They might have detected high traffic from my 4 servers and acted to block those who used those specific 4 servers. That is my assumption at the moment.
Current proposal: Members of the community can host their own bypass server via https://github.com/acheong08/ChatGPT-Proxy-V4 for smaller scale use rather than a single public proxy with high traffic.
Take this advice with caution and try to make sure you don't use accounts with money in them. We'll see what happens after a while.
Current proposal: Members of the community can host their own bypass server via https://github.com/acheong08/ChatGPT-Proxy-V4 for smaller scale use rather than a single public proxy with high traffic.
Take this advice with caution and try to make sure you don't use accounts with money in them. We'll see what happens after a while.
Unfortunately, this solution requires a plus paid account
Unfortunately, this solution requires a plus paid account
My hope is that those with plus accounts can share third party bypass servers. Perhaps charging a minuscule fee for it
I will continue running my public servers and swapping IP addresses every few days to try to avoid bans but with such high usage, it is bound to get easily detected
What differece in V1 V3 V4?
V1: only reverse ChatGPT?
V3: only reverse API
V4: only for plus user on GPT4.0?
Dose it right?
V1 - chat.openai.com
V3 - Official API
There is no V4. V4 is proxy version which is the only proxy version working
I just used v1, accessed by account and password, and talked to me a total of 5 times, and the result was that I was prompted "{"error": "Rate limited by proxy server"} (code: 429)".
I am currently using Alibaba Cloud US server, and I do not use a proxy, and I directly use the public IP address to call v1. I haven't logged in to the ChatGPT page to see if my account has been banned.
I just used v1, accessed by account and password, and talked to me a total of 5 times, and the result was that I was prompted "{"error": "Rate limited by proxy server"} (code: 429)". I am currently using Alibaba Cloud US server, and I do not use a proxy, and I directly use the public IP address to call v1. I haven't logged in to the ChatGPT page to see if my account has been banned.
I tried to log in and use chatGPT on the OpenAI website, and found that I was still able to log in normally, and my account was not banned. Seriously, when I called the API yesterday using the v1 version and displayed the restriction information, I was shocked and really thought that my account was blocked.
do not use any other chatgpt ui ,if you want to not be baned by the openai offical
Why the v1 accounts deactivated warning disappear in readme? Is the issue fixed？
Accounts no longer getting deactivated. Swapped server approach (3rd time now. Not open source yet). Will add again if the issue flares up
A WeChat group was set up for ChatGPT developers to discuss the following issues: 1. Openai api was blocked for unknown reasons; 2. How to choose an overseas server; 3. How to improve the access speed of domestic users; 4. Discussion on monetization related experience. Wait, for developers of the ChatGPT api ecosystem:
(搭建了一个微信群，面向ChatGPT开发者，探讨以下问题：1、openai api不明原因被封的；2、如何选择海外服务器；3、如何提高国内用户访问速度；4、变现相关经验探讨。等等，面向ChatGPT api生态的开发者：)
This issue is stale because it has been open for 2 days with no activity.
So there is no need to use https://github.com/acheong08/ChatGPT-Proxy-V4 right? I see there are some issues in that repo.
This issue is stale because it has been open for 2 days with no activity.
@acheong08
My plus account was deactivated few hours ago.Using V1 and GPT-4 Model.
I have used it for about 17 days with username/password auth, and change to access token auth 2 days ago, due to #1236 .
I am not sure if this is the reason causing the deactivated problem.
I am not sure if this is the reason causing the deactivated problem.
Probably not. It always uses access tokens. Even when you input email/password, all that does is fetch access token.
On another note, just checking the response codes from the servers, there doesn't seem to be anyone else getting banned. Will look into it.