README_EN.md

ChatGPT Web Midjourney Proxy

中文 | Русский язык | Français | 한국어 | Tiếng Việt | Türkçe

Disclaimer

• This project is only released on GitHub under the MIT license, free and open-source for learning purposes. There will be no form of account selling, paid services, discussion groups, etc. Beware of scams.
• This open-source project is developed based on ChenZhaoYu; it uses the midjourney API provided by midjourney-proxy and [Suno-API] as the backend.

Supported Features

• Support for the standalone Suno module, with adjustable lyrics and music style
• All original ChatGPT web features
• ChatGPT web supports custom API keys and base_url
• Midjourney text-to-image
• Midjourney blend image + text-to-image
• Midjourney image variations U1 to U4, V1 to V4, and redrawing
• Midjourney partial redrawing
• Midjourney 1.5x and 2x zoom
• Midjourney 2x and 4x HD
• Midjourney extensions to the left, right, top, and bottom
• Midjourney supports both midjourney-proxy and midjourney-proxy-plus
• Midjourney text from images
• Local image storage using localforage
• Support for different bots like midjourney and niji
• Support for InsightFace face replacement
• Midjourney image blending
• Midjourney get seed
• DALL-E-3 image generation
• ChatGPT front-end model selection
• ChatGPT front-end supports custom models, context number, and reply count
• ChatGPT supports image uploads for GPT-4-Vision-Preview
• ChatGPT supports file backend uploads (for models like GPT-4-All, GPT-4-Gizmo-xxx) by default off, can be enabled with environment variable API_UPLOADER=1
• ChatGPT supports reverse models like GPT-4-All, GPT-4-V, GPT-4-Gizmo-(gizmo_id)
• ChatGPT supports hyperlink model switching https://vercel.ddaiai.com/#/m/gpt-4-all https://vercel.ddaiai.com/#/m/gpt-4-gizmo-g-2fkFE8rbu
• ChatGPT supports ChatGPT-style hyperlink model switching https://chat.openai.com/g/g-2fkFE8rbu modified to https://vercel.ddaiai.com/#/g/g-2fkFE8rbu
• ChatGPT supports GPTs multimodal
• ChatGPT supports TTS whisper
• Instant voice recognition (browser's built-in ASR) v2.15.7 and above
• Support hyperlink change settings, suitable for one-api and new-api deployments https://vercel.ddaiai.com/#/s/t?OPENAI_API_BASE_URL=https://abc.com&OPENAI_API_KEY=sk-xxxxx&MJ_SERVER=https://abc.com&MJ_API_SECRET=sk-xxx&UPLOADER_URL=
• Support one-api and new-api deployments https://vercel.ddaiai.com/#/?settings={%22key%22:%22sk-abc%22,%22url%22:%22https://www.abc.com%22} (v.2.14.3)

Serverless - Personal Desktop Installation

• Please download the latest version from https://github.com/Dooy/chatgpt-web-midjourney-proxy/releases (choose the version suitable for your operating system)
• Choose an appropriate proxy service (preferably one that supports gpt, gpts, midjourney, claude, suno)
• Recommended proxy service https://www.openai-hk.com a key and api interface address support gpt, midjourney, claude, suno simultaneously, mj-fast as low as 0.12 RMB/image

Vercel One-Click Deployment

env Environment Variables

Environment Variable	Description	Default Value	Docker Deployment	Vercel Deployment
OPENAI_API_BASE_URL	OpenAI API interface address	https://api.openai.com	✅	✅
OPENAI_API_KEY	OpenAI API key	sk-xxxxx	✅	✅
OPENAI_API_MODEL	Default model	gpt-3.5-turbo	✅	✅
MJ_SERVER	MJ proxy interface address	Reference for setup	✅	✅
MJ_API_SECRET	MJ proxy secret	Empty	✅	✅
SUNO_SERVER	SUNO API interface address	Reference for setup	✅	✅
SUNO_KEY	SUNO API key	Empty	✅	✅
AUTH_SECRET_KEY	Access authorization password	None	✅	x
API_UPLOADER	Support upload	Disabled	✅	x
HIDE_SERVER	Hide server UI on the front end		✅	x
CUSTOM_MODELS	Custom selectable models	None	✅	✅
TJ_BAIDU_ID	Baidu Analytics ID	None	✅	✅
TJ_GOOGLE_ID	Google Analytics ID	None	✅	✅
SYS_NOTIFY	System notifications, supports HTML	None	✅	✅
DISABLE_GPT4	Disable GPT-4	None	✅	✅
GPT_URL	Custom GPT_URL=/gpts.json	None or your external link	✅	✅
UPLOAD_IMG_SIZE	GPT4V upload image size	1	✅	✅
SYS_THEME	Default theme light or dark	dark	✅	✅
MJ_IMG_WSRV	Enable wsrv image bed	None (disabled)	✅	✅
AUTH_SECRET_ERROR_COUNT	Brute force prevention: Number of verification attempts NGINX please set proxy_set_header X-Forwarded-For $remote_addr	None	✅	x
AUTH_SECRET_ERROR_TIME	Brute force prevention: Wait time in minutes	None	✅	x
CLOSE_MD_PREVIEW	Do not close input preview	None	✅	✅
UPLOAD_TYPE	Specify upload method [R2 R2 upload] [API Follow UI front-end relay] [Container Local container] [MyUrl Custom link]	Empty	✅	x
MENU_DISABLE	Disable menu options: gpts, draws, gallery, music	Empty	✅	✅
VISION_MODEL	Default recognition model Options: gpt-4o, gpt-4-turb, gpt-4-vision-preview, etc.	Empty	✅	✅
SYSTEM_MESSAGE	Custom default role message	Empty	✅	✅
CUSTOM_VISION_MODELS	Custom vision models separated by ,	Empty	✅	✅

Docker Deployment

• Requires midjourney-proxy support
• Requires [Suno-API] support

docker run --name chatgpt-web-midjourney-proxy  -d -p 6015:3002 \
-e OPENAI_API_KEY=sk-xxxxx \
-e OPENAI_API_BASE_URL=https://api.openai.com  \
-e MJ_SERVER=https://your-mj-server:6013  \
-e MJ_API_SECRET=your-mj-api-secret  \
-e SUNO_SERVER=https://your-suno-server:8000  \
-e SUNO_KEY=you-suno-key  ydlhero/chatg

pt-web-midjourney-proxy

Access http://ip:6015

File Upload:

docker run --name chatgpt-web-midjourney-proxy  -d -p 6015:3002 \
-e OPENAI_API_KEY=sk-xxxxx \
-e OPENAI_API_BASE_URL=https://api.openai.com  \
-e MJ_SERVER=https://172.17.0.1:6013  \
-e API_UPLOADER=1  -v /data/uploads:/app/uploads \
-e MJ_API_SECRET=abc123456  ydlhero/chatgpt-web-midjourney-proxy

If the front-end UI sets OPENAI_API_KEY and OPENAI_API_BASE_URL, the image upload will also follow the OPENAI_API_BASE_URL.

curl -X POST -H "Content-Type: multipart/form-data" -F "file=@/path/to/file" http://OPENAI_API_BASE_URL/v1/upload

Return format

{
"url":"https://xxxxxxx.jpg"
}

Midjourney-proxy API Docker Deployment

For more references, go to midjourney-proxy open-source project.

docker run -d --name mj6013  -p 6013:8080  \
-e mj.discord.guild-id=Discord Server ID  \
-e mj.discord.channel-id=Discord Server Group ID   \
-e mj.queue.timeout-minutes=6 \
-e mj.api-secret=abc123456 \
-e mj.discord.user-token=**********  \
--restart=always novicezk/midjourney-proxy:2.5.5

More Demonstrations

Custom Server API Key, Base_URL:

GPTS GTP Store

Suno Music Creation

Recording Whisper and TTS

Partial Redraw:

Face Replacement

Image Blending

Support for Image Upload for GPT-4-Vision-Preview

Mobile:

File Upload Support for Cloudflare R2 Storage

• Cloudflare R2 storage 10 GB/month free https://www.cloudflare.com/zh-cn/developer-platform/r2/
• Configuration reference https://zhuanlan.zhihu.com/p/658058503
• Vercel does not support R2 storage

R2_DOMAIN=
R2_BUCKET_NAME=
R2_ACCOUNT_ID=
R2_KEY_ID=
R2_KEY_SECRET=

File Server Request Priority

R2 > Front-end UI set file service > Backend file service > Follow proxy

Brute Force Prevention Settings

• Vercel is not supported; only Docker deployment is supported
• If NGINX is mounted in front, configure proxy_set_header X-Forwarded-For $remote_addr;
• Parameters: 3 failed verification attempts, can only reattempt after 10 minutes

# Secret key Note: Only alphanumeric characters
AUTH_SECRET_KEY=my888god
# Brute force: verification attempts Note: Number; NGINX please set proxy_set_header X-Forwarded-For $remote_addr;
AUTH_SECRET_ERROR_COUNT=3
# Brute force: wait time in minutes Note: Number
AUTH_SECRET_ERROR_TIME=10

• Script:

docker run --name chatgpt-web-midjourney-proxy  -d -p 6015:3002 \
-e OPENAI_API_KEY=sk-xxxxx \
-e OPENAI_API_BASE_URL=https://api.openai.com  \
-e MJ_SERVER=https://172.17.0.1:6013  \
-e MJ_API_SECRET=abc123456 \
-e API_UPLOADER=1  -v /data/uploads:/app/uploads \
-e AUTH_SECRET_KEY=your-english-password -e AUTH_SECRET_ERROR_COUNT=3 \
-e AUTH_SECRET_ERROR_TIME=10 ydlhero/chatgpt-web-midjourney-proxy

License

MIT © Dooy

Others

If you find this project helpful, please consider giving it a star or donating to us.

Donations

If my open-source project is helpful to you, please consider sponsoring me through any of the following methods:
Payment remarks with your contact information

WeChat Donation

Alipay Donation