Turn WeChat into a programmable AI assistant
Deploy in one click, scan to go live
Every user gets a dedicated cloud host that runs a real WeChat desktop client and exposes a full HTTP / WebSocket API. Call the API to send and receive messages directly, or plug in Claude Code and let an AI agent read and reply to your WeChat conversations for you.
Core capabilities
It drives a real WeChat desktop client — behaving close to a human user — and exposes a unified API covering the full flow from login and messaging to event streaming and chat history.
Send messages
Send text, images, files and video to a contact or group via /api/messages/send, with the result returned in real time.
Real-time listening
Subscribe to inbound message events over the WebSocket /ws and push new messages into your ticketing, CRM or alerting systems instantly.
History & search
Fetch and refresh conversation history and search by keyword — ideal for message retention, QA review and analytics.
Contacts & sync
Sync contacts and conversation lists and fetch the latest state on demand, cutting maintenance overhead on your side.
Status & health
/api/status and /ping expose online status and availability, so you can wire up monitoring and auto-recovery.
Dedicated auth & ops
Each instance has its own API_TOKEN, plus QR login, logout, restart and RDP remote desktop for troubleshooting.
Let Claude Code run your WeChat
WisdomChannel is a Claude Code channel plugin that bridges your WeChat instance with Claude Code: incoming WeChat messages are handed to Claude, and its replies are sent back — turning your WeChat into an AI assistant that does the work for you.
- Direct messages are forwarded automatically; group chats only trigger on
@you - Built-in allowlist access control — only respond to contacts you approve
- Two modes: an interactive MCP channel, or the unattended
wisdom-channel bridgeauto-reply loop - One-line install:
pip install wisdom-channel(requires Claude Code CLI v2.1.80+)
pip install wisdom-channelpoint it at your instance + token
setx WISDOM_API_URL "http://<your-instance-ip>:8000" setx WISDOM_API_TOKEN "<API_TOKEN>"
bridge: new message -> Claude -> auto-reply
wisdom-channel bridge --model sonnet
Quick start
Each instance has its own API address and token (API_TOKEN). Call the instance's own :8000 endpoint to send a message.
curl -X POST "http://<your-instance-ip>:8000/api/messages/send" \
-H "Authorization: Bearer <API_TOKEN>" \
-H "Content-Type: application/json" \
-d '{
"target": "Alice",
"type": "text",
"text": "Hello from WeChat Bot"
}'
Live in four steps
A standardized deployment flow that lowers the integration and ops barrier.
Create the app
Create a WeChat Bot app on the platform and a dedicated cloud host is provisioned for you automatically.
Scan to log in
Scan the QR with your phone's WeChat in the console, then grab the instance API address, API_TOKEN and RDP entry.
Call the API or add Claude
Call the :8000 endpoints directly, or install wisdom-channel to let Claude Code auto-reply.
Connect your systems
Wire message events into your internal flows and build an observable, maintainable automation workflow.
Use cases
For teams and individuals that want stable WeChat capabilities with an AI automation layer on top.
- AI customer support with 24/7 coverage (Claude Code)
- Private-domain operations and scheduled outreach
- Notification bots and alert push
- Support message routing and ticket integration
- Chat history retention, search and QA analytics
- Personal WeChat assistant: triage, remind and reply
FAQ
How does it connect to WeChat?
Each instance runs a real WeChat desktop client, and Wisdom drives it through desktop automation to send and receive messages — behaving close to a human user, not a third-party web protocol.
How do I make the AI reply automatically?
Install wisdom-channel and connect Claude Code. Direct messages are forwarded automatically, groups trigger only on @you, and an allowlist controls who gets a response.
Does it support real-time events?
Yes. Subscribe to inbound message events over the WebSocket /ws for real-time processing and system integration.
How do I troubleshoot and operate it?
Log into the host over RDP remote desktop, and use the /api/status health check plus logout/restart endpoints to diagnose and recover quickly.
Any usage guidelines?
Follow WeChat's platform rules and keep sending frequency and content reasonable to avoid impacting normal accounts.
Get started with WeChat Bot
A unified deployment, auth and ops model — plus Claude Code — to ship your WeChat automation fast.