midjourney-proxy

English | 中文

Proxy the Discord channel for MidJourney to enable API-based calls for AI drawing

Main Functions

Supports Imagine instructions and related actions
Supports adding image base64 as a placeholder when using the Imagine command
Supports Blend (image blending) and Describe (image to text) commands
Supports real-time progress tracking of tasks
Supports translation of Chinese prompts, requires configuration of Baidu Translate or GPT
Prompt sensitive word pre-detection, supports override adjustment
User-token connects to WSS (WebSocket Secure), allowing access to error messages and full functionality
Supports multi-account configuration, with each account able to set up corresponding task queues

🚀 For more features, please refer to midjourney-proxy-plus

Supports all the features of the open-source version

Supports Shorten (prompt analysis) command

Supports focus shifting: Pan ⬅️ ➡️ ⬆️ ⬇️

Supports image zooming: Zoom 🔍

Supports local redrawing: Vary (Region) 🖌

Supports nearly all associated button actions and the 🎛️ Remix mode

Supports retrieving the seed value of images

Account pool persistence, dynamic maintenance

Supports retrieving account /info and /settings information

Account settings configuration

Supports Niji bot robot

Supports InsightFace face replacement robot

Embedded management dashboard page

Prerequisites for use

Register and subscribe to MidJourney, create your own server and channel, refer to https://docs.midjourney.com/docs/quick-start
Obtain user Token, server ID, channel ID: Method of acquisition

Quick Start

Railway: Based on the Railway platform, no need for your own server: Deployment method ; If Railway is not available, you can start using Zeabur instead.
Zeabur: Based on the Zeabur platform, no need for your own server: Deployment method
Docker: Start using Docker on a server or locally: Deployment method

Local development

Depends on Java 17 and Maven
Change configuration items: Edit src/main/resources/application.yml
Project execution: Start the main function of ProxyApplication
After changing the code, build the image: Uncomment VOLUME in the Dockerfile, then execute docker build . -t midjourney-proxy

Configuration items

mj.accounts: Refer to Account pool configuration
mj.task-store.type: Task storage method, default is in_memory (in memory, lost after restart), Redis is an alternative option.
mj.task-store.timeout: Task storage expiration time, tasks are deleted after expiration, default is 30 days.
mj.api-secret: API key, if left empty, authentication is not enabled; when calling the API, you need to add the request header 'mj-api-secret'.
mj.translate-way: The method for translating Chinese prompts into English, options include null (default), Baidu, or GPT.
For more configuration options, see Configuration items

Precautions

Frequent image generation and similar behaviors may trigger warnings on your Midjourney account. Please use with caution.
For common issues and solutions, see Wiki / FAQ
Interested friends are also welcome to join the discussion group. If the group is full from scanning the code, you can add the administrator’s WeChat to be invited into the group. Please remark: mj join group.

Application Project

If you have a project that depends on this one and is open source, feel free to contact the author to be added here for display.

wechat-midjourney : A proxy WeChat client that connects to MidJourney, intended only as an example application scenario, will no longer be updated.
chatgpt-web-midjourney-proxy : chatgpt web, midjourney, gpts,tts, whisper A complete UI solution
stable-diffusion-mobileui : SDUI, based on this interface and SD (System Design), can be packaged with one click to generate H5 and mini-programs.
MidJourney-Web : 🍎 Supercharged Experience For MidJourney On Web UI

Open API

Provides unofficial MJ/SD open API, add administrator WeChat for inquiries, please remark: api

Others

If you find this project helpful, please consider giving it a star.

JorWang/midjourney-proxy