README translated to English

README.md

@@ -32,9 +32,9 @@
 > - [x] 免登录 GPT-3.5 对话
 > - [x] GPT-3.5 模型对话（传入模型名不包含 gpt-4，则默认使用 gpt-3.5，也就是 text-davinci-002-render-sha）
 > - [x] GPT-4 系列模型对话（传入模型名包含: gpt-4，gpt-4o，gpt-4o-mini，gpt-4-moblie 即可使用对应模型，需传入 AccessToken）
-> - [x] O1 系列模型对话（传入模型名包含 o1-preview，o1-mini 即可使用对应模型，需传入 AccessToken）
+> - [x] O1/O3/O4 系列模型对话（传入模型名包含 o3，o4-mini等 即可使用对应模型，需传入 AccessToken）
 > - [x] GPT-4 模型画图、代码、联网
-> - [x] 支持 GPTs（传入模型名：gpt-4-gizmo-g-*）
+> - [x] 支持 GPTs（传入模型名：gpt-4-gizmo-g-*，前面模型可改，Team工作区项目需要使用这个）
 > - [x] 支持 Team Plus 账号（需传入 team account id）
 > - [x] 上传图片、文件（格式为 API 对应格式，支持 URL 和 base64）
 > - [x] 可作为网关使用，可多机分布部署
@@ -92,6 +92,13 @@ curl --location 'http://127.0.0.1:5005/v1/chat/completions' \
 
 > - `AccessToken` 获取: chatgpt官网登录后，再打开 [https://chatgpt.com/api/auth/session](https://chatgpt.com/api/auth/session) 获取 `accessToken` 这个值。
 > - `RefreshToken` 获取: 此处不提供获取方法。
+> - `ChatGPT-Account-ID` 获取:
+>   - 方法①，访问 <https://chatgpt.com/admin> 使用F12找到请求 <https://chatgpt.com/backend-api/accounts/UUID/users>，这个ID就是。
+>   - 方法②，还是上一步的地方读取响应中的`account_user_id`，后面的那个UUID。
+>   - 方法③，访问 <https://chatgpt.com/api/auth/session> 找到 account 下面的 id（不是 organizationId），或者在网页发起工作区对话请求时F12请求头就有这个。
+> - `gizmo` ID 获取:
+>   - 方法①，打开项目包，URL中的g-p开头，去掉后面的项目名称英文版的部分就是。
+>   - 方法②，在工作区项目进行对话，在请求体中找到 `gizmo_id`。
 > - 免登录 gpt-3.5 无需传入 Token。
 
 ## Tokens 管理

README_EN.md

@@ -0,0 +1,211 @@
+# CHAT2API
+
+🤖 A Simple ChatGPT TO API Proxy
+
+🌟 Free and Unlimited `GPT-3.5` Without Account
+
+💥 Supports AccessToken for Account Usage, Supports `O3-mini/high`, `O1/mini/Pro`, `GPT-4/4o/mini`, `GPTs`
+
+🔍 Response Format Completely Consistent with Real API, Compatible with Almost All Clients
+
+👮 Companion User Management Platform [Chat-Share](https://github.com/h88782481/Chat-Share) Requires Environment Variables to be Configured in Advance (Set ENABLE_GATEWAY to True, AUTO_SEED to False)
+
+## Community Group
+
+[https://t.me/chat2api](https://t.me/chat2api)
+
+Before Asking Questions, Please Read the Repository Documentation Thoroughly, Especially the FAQ Section.
+
+When Asking Questions, Please Provide:
+
+1. Startup Log Screenshot (Sensitive Information Masked, Including Environment Variables and Version Number)
+2. Error Log Information (Sensitive Information Masked)
+3. Status Code and Response Body of the Interface Return
+
+## Features
+
+### Latest Version Number Stored in `version.txt`
+
+### Reverse API Features
+> - [x] Streaming and Non-streaming Transmission
+> - [x] Login-free GPT-3.5 Conversation
+> - [x] GPT-3.5 Model Conversation (Default to GPT-3.5 if model name does not contain gpt-4, i.e., text-davinci-002-render-sha)
+> - [x] GPT-4 Series Model Conversation (Use corresponding model by including: gpt-4, gpt-4o, gpt-4o-mini, gpt-4-moblie, requires AccessToken)
+> - [x] O1/O3/O4 series model dialogue (the corresponding model can be used by passing in model names including o3, o4-mini, etc., and AccessToken needs to be passed in)
+> - [x] GPT-4 Model Drawing, Code, Internet Access
+> - [x] Support GPTs (input model name: gpt-4-gizmo-g-*, the previous model can be changed, Team workspace projects need to use this)
+> - [x] Support for Team Plus Accounts (Requires team account id)
+> - [x] Upload Images and Files (API-compatible format, supports URL and base64)
+> - [x] Can be Used as a Gateway, Supports Multi-machine Distributed Deployment
+> - [x] Multi-account Polling, Supporting Both `AccessToken` and `RefreshToken`
+> - [x] Request Failure Retry, Automatic Token Polling
+> - [x] Tokens Management, Support Upload and Clearing
+> - [x] Periodic Refresh of `AccessToken` using `RefreshToken` / Full Non-forced Refresh on Startup, Forced Refresh at 3 AM Every 4 Days
+> - [x] Support File Download, Requires History Record Enabled
+> - [x] Supports Model Inference Process Output for `O3-mini/high`, `O1/mini/Pro`
+
+### Official Website Mirror Features
+> - [x] Support for Official Native Mirror
+> - [x] Backend Account Pool Random Selection, `Seed` Setting for Random Accounts
+> - [x] Direct Login with `RefreshToken` or `AccessToken`
+> - [x] Supports `O3-mini/high`, `O1/mini/Pro`, `GPT-4/4o/mini`
+> - [x] Sensitive Information Interface Disabled, Partial Setting Interfaces Disabled
+> - [x] /login Page, Automatic Redirect to Login Page After Logout
+> - [x] /?token=xxx Direct Login, xxx is `RefreshToken` or `AccessToken` or `SeedToken` (Random Seed)
+> - [x] Supports Session Isolation with Different SeedTokens
+> - [x] Supports `GPTs` Store
+> - [x] Supports Official-exclusive Features like `DeepReaserch`, `Canvas`
+> - [x] Supports Language Switching
+
+> TODO
+> - [ ] None, Welcome to Submit `Issues`
+
+## Reverse API
+
+A fully `OpenAI`-compatible API that supports passing `AccessToken` or `RefreshToken`, usable with GPT-4, GPT-4o, GPT-4o-Mini, GPTs, O1-Pro, O1, O1-Mini, O3-Mini, O3-Mini-High:
+
+```bash
+curl --location 'http://127.0.0.1:5005/v1/chat/completions' \
+--header 'Content-Type: application/json' \
+--header 'Authorization: Bearer {{Token}}' \
+--data '{
+     "model": "gpt-3.5-turbo",
+     "messages": [{"role": "user", "content": "Say this is a test!"}],
+     "stream": true
+   }'
+```
+
+Pass your account's `AccessToken` or `RefreshToken` as `{{ Token }}`.
+You can also fill in the value of the environment variable `Authorization`, which will randomly select a backend account.
+
+If you have a team account, you can pass in `ChatGPT-Account-ID` to use the Team workspace:
+
+- Method One:
+  Pass the `ChatGPT-Account-ID` value in the `headers`
+- Method Two:
+  `Authorization: Bearer <AccessToken or RefreshToken>,<ChatGPT-Account-ID>`
+
+If the `AUTHORIZATION` environment variable is set, you can use the set value as `{{ Token }}` for multi-Token polling.
+
+> - `AccessToken` retrieval: Log in to the ChatGPT website, then open [https://chatgpt.com/api/auth/session](https://chatgpt.com/api/auth/session) to get the `accessToken` value.
+> - `RefreshToken` retrieval: Method not provided here.
+> - Get `ChatGPT-Account-ID`:
+> - Method ①, visit <https://chatgpt.com/admin> and use F12 to find the request <https://chatgpt.com/backend-api/accounts/UUID/users>, this ID is it.
+> - Method ②, read the `account_user_id` in the response from the previous step, the UUID behind it.
+> - Method ③, visit <https://chatgpt.com/api/auth/session> and find the id under account (not organizationId), or when the web page initiates a workspace conversation request, the F12 request header has this.
+> - Get `gizmo` ID:
+> - Method ①, open the project package, the URL starts with g-p, and remove the English version of the project name behind it.
+> - Method ②, have a conversation in the workspace project, and find `gizmo_id` in the request body.
+> - No login required for gpt-3.5, no token needed.
+
+## Tokens Management
+
+1. Configure the environment variable `AUTHORIZATION` as an `authorization code`, then run the program.
+2. Access `/tokens` or `/{api_prefix}/tokens` to view the current number of Tokens, upload new Tokens, or clear Tokens.
+3. Pass the `authorization code` configured in `AUTHORIZATION` to use polling Tokens for conversation
+
+![tokens.png](docs/tokens.png)
+
+## Official Website Native Mirror
+
+1. Set the environment variable `ENABLE_GATEWAY` to `true`, then run the program. Note that enabling this allows others to directly access your gateway via domain.
+2. Upload `RefreshToken` or `AccessToken` on the Tokens management page
+3. Access `/login` to reach the login page
+
+![login.png](docs/login.png)
+
+4. Use on the official native mirror page
+
+![chatgpt.png](docs/chatgpt.png)
+
+## Environment Variables
+
+Each environment variable has a default value. If you don't understand the meaning of an environment variable, do not set it, and definitely do not pass an empty value. Strings do not require quotes.
+
+| Category | Variable Name | Example Value | Default Value | Description |
+| -------- | ------------- | ------------- | ------------- | ----------- |
+| Security | API_PREFIX | `your_prefix` | `None` | API prefix password. Without setting, it's easily accessible. After setting, request `/your_prefix/v1/chat/completions` |
+| | AUTHORIZATION | `your_first_authorization`,`your_second_authorization` | `[]` | Authorization codes you set for using multiple accounts in token rotation, separated by English commas |
+| | AUTH_KEY | `your_auth_key` | `None` | Set this for private gateway requiring `auth_key` request header |
+| Request | CHATGPT_BASE_URL | `https://chatgpt.com` | `https://chatgpt.com` | ChatGPT gateway address. Setting this will change the requested website. Multiple gateways can be separated by commas |
+| | PROXY_URL | `http://ip:port`,`http://username:password@ip:port` | `[]` | Global proxy URL, enabled when encountering 403, multiple proxies separated by commas |
+| | EXPORT_PROXY_URL | `http://ip:port` or `http://username:password@ip:port` | `None` | Export proxy URL to prevent source site IP leakage when requesting images and files |
+| Functionality | HISTORY_DISABLED | `true` | `true` | Whether to not save chat history and return conversation_id |
+| | POW_DIFFICULTY | `00003a` | `00003a` | Proof of Work difficulty to solve, do not set if unsure |
+| | RETRY_TIMES | `3` | `3` | Number of error retries. Using `AUTHORIZATION` will automatically randomly/round-robin to the next account |
+| | CONVERSATION_ONLY | `false` | `false` | Whether to directly use the conversation interface. Enable only if your gateway supports automatic POW resolution |
+| | ENABLE_LIMIT | `true` | `true` | When enabled, it will not attempt to bypass official request limits to minimize account suspension risk |
+| | UPLOAD_BY_URL | `false` | `false` | When enabled, dialogue follows `URL+space+text`, automatically parsing and uploading URL content. Multiple URLs separated by spaces |
+| | SCHEDULED_REFRESH | `false` | `false` | Whether to periodically refresh `AccessToken`. When enabled, all will be non-forcefully refreshed on startup, and forcefully refreshed at 3 AM every 4 days |
+| | RANDOM_TOKEN | `true` | `true` | Whether to randomly select backend `Token`. Enabled for random backend account selection, disabled for sequential round-robin |
+| Gateway | ENABLE_GATEWAY | `false` | `false` | Whether to enable gateway mode. When enabled, mirror sites can be used, but defenses will be lowered |
+| | AUTO_SEED | `false` | `true` | Whether to enable random account mode. Enabled by default. After inputting `seed`, randomly match backend `Token`. Disable to manually interface for `Token` management |
+
+## Deployment
+
+### Zeabur Deployment
+
+[![Deploy on Zeabur](https://zeabur.com/button.svg)](https://zeabur.com/templates/6HEGIZ?referralCode=LanQian528)
+
+### Direct Deployment
+
+```bash
+git clone https://github.com/LanQian528/chat2api
+cd chat2api
+pip install -r requirements.txt
+python app.py
+```
+
+### Docker Deployment
+
+You need to install Docker and Docker Compose.
+
+```bash
+docker run -d \
+  --name chat2api \
+  -p 5005:5005 \
+  lanqian528/chat2api:latest
+```
+
+### (Recommended, PLUS account compatible) Docker Compose Deployment
+
+Create a new directory, for example, chat2api, and enter it:
+
+```bash
+mkdir chat2api
+cd chat2api
+```
+
+Download the docker-compose.yml file from the repository in this directory:
+
+```bash
+wget https://raw.githubusercontent.com/LanQian528/chat2api/main/docker-compose-warp.yml
+```
+
+Modify the environment variables in the docker-compose-warp.yml file, then save:
+
+```bash
+docker-compose up -d
+```
+
+## Common Issues
+
+> - Error Codes:
+>   - `401`: Current IP does not support login-free access. Try changing IP address, setting proxy in `PROXY_URL` environment variable, or authentication failure.
+>   - `403`: Check specific error message in logs.
+>   - `429`: Current IP has exceeded hourly request limit. Wait and retry, or change IP.
+>   - `500`: Internal server error, request failed.
+>   - `502`: Server gateway error or network unavailable. Try changing network environment.
+
+> - Known Situations:
+>   - Many Japanese IPs do not support login-free access. Recommend using US IP for free GPT-3.5.
+>   - 99% of accounts support free `GPT-4o`, but activation depends on IP region. Japan and Singapore IPs currently have higher activation probability.
+
+> - What is the `AUTHORIZATION` environment variable?
+>   - A self-set authentication for chat2api. Setting this allows using saved Tokens in rotation. Pass it as `APIKEY` during requests.
+> - How to obtain AccessToken?
+>   - Log in to ChatGPT official website, then open [https://chatgpt.com/api/auth/session](https://chatgpt.com/api/auth/session) to get the `accessToken` value.
+
+## License
+
+MIT License

api/models.py

@@ -8,11 +8,14 @@
     "gpt-4-turbo": "gpt-4-turbo-2024-04-09",
     "gpt-4o": "gpt-4o-2024-08-06",
     "gpt-4o-mini": "gpt-4o-mini-2024-07-18",
+    "gpt-4-5": "gpt-4-5",
     "o1-preview": "o1-preview-2024-09-12",
     "o1-mini": "o1-mini-2024-09-12",
     "o1": "o1-2024-12-18",
     "o3-mini": "o3-mini-2025-01-31",
     "o3-mini-high": "o3-mini-high-2025-01-31",
+    "o4-mini": "o4-mini-2025-04-16",
+    "o4-mini-high": "o4-mini-high-2025-04-16",
     "claude-3-opus": "claude-3-opus-20240229",
     "claude-3-sonnet": "claude-3-sonnet-20240229",
     "claude-3-haiku": "claude-3-haiku-20240307",