whatisgithub

What is usechat?

cosmosshadow/usechat — explained in plain English

Analysis updated 2026-05-18

37TypeScriptAudience · developerComplexity · 4/5Setup · hard

In one sentence

A local tool that lets an AI agent operate the WeChat desktop app on your computer through screenshots and OCR, without hacking protocols or reading your chat data.

Mindmap

mindmap
  root((UseChat))
    What it does
      Reads WeChat messages
      Sends messages and files
      Watches conversations live
    Tech stack
      TypeScript
      Node.js CLI
      Vision AI model
    Use cases
      AI agent WeChat access
      Automated replies
      Message monitoring
    Audience
      Developers
      Agent builders

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Let an AI agent read the latest messages in a WeChat chat as JSON or Markdown.

USE CASE 2

Have an agent send a WeChat reply once a human approves the text.

USE CASE 3

Watch a specific WeChat conversation and get notified as new messages arrive.

USE CASE 4

Plug UseChat into a custom agent runtime as a stdio tool so it can read and write messages programmatically.

What is it built with?

TypeScriptNode.jsnpmOpenAI-compatible API

How does it compare?

cosmosshadow/usechathotakus/opencode-visual-cacheimgoodbai/mapgogogo
Stars373737
LanguageTypeScriptTypeScriptTypeScript
Setup difficultyhardeasymoderate
Complexity4/52/53/5
Audiencedeveloperdevelopergeneral

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · hard Time to first run · 1h+

Requires installing a separate Helper app, granting OS-level screen and input permissions, and configuring a vision-capable API key.

So what is it?

UseChat is a tool that lets an AI agent, like Claude Code, Cursor, Codex, or OpenCode, operate the WeChat desktop app on your own computer, the same way a person would sit in front of it. It does this by taking screenshots of the WeChat window, reading the text with OCR, and understanding the layout, then turning all of that into commands the agent can read, write, and check. It does not break into WeChat's private protocols, does not inject code into the client, does not read your local chat database, and does not take over your account. Everything happens through screenshots, mouse clicks, keyboard input, and clipboard actions that the operating system already allows. The first supported app is WeChat's desktop client, on both macOS and Windows. With the command line tool, you can read the messages currently visible in a chat, send text, files, images, or videos, and watch a conversation for new messages as they arrive, with results shown as readable text or as JSON for machines. There is also a serve mode that keeps a steady connection open so an external agent runtime can send read and write requests to it directly. Reading messages requires a vision capable AI model, since the tool has to look at a screenshot and turn it into a structured list of messages. Sending messages does not use a model at all. Setup involves installing the command line package through npm, plus a separate Helper program that actually handles the screen capture, clicks, and OCR on your machine, since the Helper is not installed automatically with the main package. UseChat comes from a company called Shennian, which built a broader WeChat automation and multi agent control product, this project is the local connector piece they open sourced separately. The tool includes built in checks, a doctor command, to confirm the Helper is installed, permissions are granted, and WeChat is logged in, along with diagnostic logs that avoid saving raw screenshots or full clipboard contents.

Copy-paste prompts

Prompt 1
Write a script using UseChat's CLI to read the last 20 messages from a WeChat chat named ABC and output them as JSON.
Prompt 2
Set up UseChat's serve --stdio mode and connect it to my custom agent so it can call the read and write tools.
Prompt 3
Configure UseChat to use an OpenAI-compatible vision model for reading WeChat messages and list the environment variables I need to set.
Prompt 4
Show me how to use UseChat's doctor command to diagnose why my WeChat Helper isn't detecting permissions.

Frequently asked questions

What is usechat?

A local tool that lets an AI agent operate the WeChat desktop app on your computer through screenshots and OCR, without hacking protocols or reading your chat data.

What language is usechat written in?

Mainly TypeScript. The stack also includes TypeScript, Node.js, npm.

How hard is usechat to set up?

Setup difficulty is rated hard, with roughly 1h+ to a first successful run.

Who is usechat for?

Mainly developer.

Open on GitHub → Ask about another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.