Connecting Meta Muse to WeChat: Using ClawBot & iLink Protocol for Tasks & Image Delivery

Who it's for

• Users actively using Meta Muse who want WeChat as a daily task dispatch and notification interface
• Anyone wanting an AI agent in personal WeChat that can invoke real tools and run shell commands
• Users who prefer not to manage cloud servers or public IPs and want Muse to self-configure the bridge

How to do it

1. Enable the ClawBot plugin in WeChat under 'Me → Settings → Plugins'
2. Send the official protocol documentation and build prompts to your Muse session
3. Wait for Muse to configure the background bridge script and task wakeup hooks in its cloud Linux environment
4. Scan the generated QR code with WeChat on your phone to authorize and link your account
5. Send an initial greeting in the WeChat ClawBot chat to verify connectivity
6. Dispatch real tool tasks from WeChat and enable direct image attachment transmission

Watch out for

• ClawBot is an experimental WeChat feature; availability and interface contracts may change over time
• Messaging capability does not equal task capability: verify that Muse configured both the polling daemon and the tool worker hook
• This workflow supports text instructions and text/image return payloads; audio, video, and direct file uploads are currently unsupported

Author: UPGPTs Team·First published: 2026-10-18·Last updated: 2026-10-18

TL;DR: By leveraging WeChat's official ClawBot plugin and the iLink protocol, you can instruct Meta Muse to self-configure a background polling daemon and task wakeup hook. After scanning an authorization QR code with your phone, you can send instructions directly in WeChat, execute real agent tasks, and receive text summaries and generated images back into your chat. No public IP or self-hosted server required — typical tasks return results within 1 to 2 minutes.

Testing Environment & Security Note: Tested in October 2026. This setup uses WeChat's official Bot endpoints (ilinkai.weixin.qq.com) authorized via standard in-app QR confirmation — no reverse-engineered protocols or memory hooking.

---

📚 Table of Contents

• Prerequisites
• Comparison: WeChat ClawBot vs Other Approaches
• Step 1: Enable the ClawBot Plugin in WeChat
• Step 2: Send Protocol Docs & Setup Prompt to Muse
• Step 3: Scan QR Code to Authorize WeChat Account
• Step 4: Test Basic Chat Connectivity
• Step 5: Test Tool Execution & Enable Direct Image Delivery
• Gateway Architecture & Core Parameters
• Troubleshooting Common Issues
• Appendix: Muse WeChat Setup Prompt Template
• Frequently Asked Questions (FAQ)
• Related Guides

---

Prerequisites

Before starting, ensure you have:

1. An active Meta Muse account: Muse is Meta's cloud-hosted personal AI agent running in a Linux sandbox. If you don't have an account, follow our Meta Muse Quick-Start & Waitlist Guide first.
2. A WeChat account with ClawBot plugin access: Open WeChat and verify under Me → Settings → Plugins. If not visible, see Troubleshooting.

---

Comparison: WeChat ClawBot vs Other Approaches

Method	Protocol / Channel	Overseas Network Needed	Image / File Returns	Security & Ban Risk	Best For
WeChat ClawBot (This Guide)	Tencent iLink Official Bot API	No (Muse handles it in cloud)	Supported (Images)	Official QR auth, High Safety	Everyday WeChat users wanting zero server maintenance
WhatsApp Official	Meta Native Channel	Yes (Phone needs proxy)	Native multimodal	Official, Zero Risk	Global users already on WhatsApp
WeCom Custom App	Enterprise WeChat API	No	Supported via app permissions	Official, Stable	Corporate teams with admin privileges
Reverse-Engineered Hook	Web/iPad protocol simulation	No	Tool dependent	Extremely High Ban Risk	Not recommended for personal numbers

To see how other open-source agents connect to WeChat, compare with our Hermes Agent WeChat Integration Guide.

---

Step 1: Enable the ClawBot Plugin in WeChat

Open WeChat on your mobile phone, navigate to Me → Settings → Plugins, locate ClawBot, and tap Enable.

Once enabled, a dedicated "WeChat ClawBot" contact appears in your chat list. This channel interfaces directly with Tencent's official bot backend via the OpenClaw standard.

---

Step 2: Send Protocol Docs & Setup Prompt to Muse

If you ask Muse to connect to personal WeChat without context, it will decline, stating personal accounts are unsupported and suggesting WhatsApp or Enterprise WeChat instead.

The solution is to provide Muse with the official Tencent open-source protocol documentation so it can implement the bridge script inside its cloud environment. (To understand how Muse operates in its cloud environment, see Registering Claude Accounts with Muse).

Provide Muse with these references:

• Tencent openclaw-weixin protocol: https://github.com/tencent/openclaw-weixin/blob/HEAD/docs/protocol_zh_CN.md
• WeChat Open Platform docs: https://developers.weixin.qq.com/doc/aispeech/knowledge/openapi/Clawbotrelated.html

Muse will create the bridge script (~/workspace/wechat-bridge/bridge.py) and register background monitoring hooks. Ensure it creates both:

1. Message Listener: A long-polling process reading incoming private chat messages and pushing them to an inbox/ directory.
2. Task Worker & Hook: A periodic scanner claiming unhandled tasks, launching tool-enabled workers, and returning responses via the API.

---

Step 3: Scan QR Code to Authorize WeChat Account

Once Muse reports the environment is ready, reply to prompt it for the login QR code.

Open WeChat's built-in "Scan" tool to scan the image:

• The QR code is valid for 5 minutes; ask Muse to regenerate if it expires.
• If prompted on mobile for a verification code, send the code to Muse to confirm.
• Once confirmed on your phone, the server issues a session token, and Muse starts the bridge daemon.

---

Step 4: Test Basic Chat Connectivity

Open the WeChat ClawBot conversation in your chat list and send a greeting message (e.g., "Hello").

The initial message establishes conversational context. Upon receiving a response, send a follow-up query to confirm bidirectional communication:

---

Step 5: Test Tool Execution & Enable Direct Image Delivery

With basic messaging verified, test whether Muse can trigger actual tools from WeChat instructions.

Send an image generation prompt in WeChat:

"Generate a photo of a young woman enjoying coffee in a café."

Muse will awaken a tool-enabled worker, generate the image, and initially store it in its cloud media repository.

Initial scripts usually only implement plain text replies, instructing you to open the Muse web UI to view the image:

Simply instruct Muse in the WeChat conversation:

"Please upgrade the bridge to upload images to WeChat and send them directly into our chat."

Muse will update its bridge logic to invoke WeChat's media upload API. Subsequent image prompts will deliver rendered photos directly into your chat:

---

Gateway Architecture & Core Parameters

Component	Technical Spec	Description
Endpoint	ilinkai.weixin.qq.com	Official Tencent bot backend; clean and compliant
Inbound Queue	35s HTTP Long-Polling	Background daemon listening for private messages into inbox/
Worker Hook	15s Local Directory Scan	Scans for unclaimed tasks and atomizes them into running/
Roundtrip Time	1 to 2 minutes	Total time = Polling (~35s) + Hook (~15s) + Worker spinup & execution
Payload Support	Text prompts & Text/Image returns	Audio, video, and inbound document parsing currently unsupported
Chat Isolation	Separated UI & Process	WeChat dialogue is isolated from Muse web chat history

---

Troubleshooting Common Issues

• Cannot find ClawBot in WeChat Plugins
  • ClawBot is in phased rollout. Upgrade WeChat to the newest build, check under Settings → General → Auxiliary Features / Plugins, or test with an alternate account.
• Messages in WeChat get no response after QR auth
  • Usually caused by a dead bridge.py process or missing worker hook. Return to the Muse web chat and instruct it to check process status and ensure wechat-inbox hooks are active.
• Why does each command take 1-2 minutes?
  • This is expected behavior due to the 35s polling window, 15s scan interval, and container cold-start time. Avoid repeating prompts within 30 seconds to prevent race conditions.
• Generated images don't appear in chat
  • Message Muse: "Add image upload functionality so generated files are sent directly to this chat via media_id."

---

Appendix: Muse WeChat Setup Prompt Template

You can copy and send this exact prompt to Muse to begin setup:

Please review these two WeChat bot protocol specifications:
1. https://github.com/tencent/openclaw-weixin/blob/HEAD/docs/protocol_zh_CN.md
2. https://developers.weixin.qq.com/doc/aispeech/knowledge/openapi/Clawbotrelated.html

Based on the Tencent Bot protocol (ilinkai.weixin.qq.com) and QR authentication:
1. Write a bridge script supporting QR login, 35s long-polling message reception, and image returns.
2. Configure a periodic hook to claim incoming tasks, trigger tool-enabled workers, and send responses to WeChat.
3. Verify your local environment and network. When ready, notify me so I can request the QR code to scan.

---

Frequently Asked Questions (FAQ)

Why does Muse refuse when I simply ask it to connect to WeChat?

Muse only has native out-of-the-box support for WhatsApp. Without documentation, it assumes personal WeChat is inaccessible. Providing the open-source openclaw-weixin protocol docs allows Muse to read the OpenAPI specs and write its own iLink bridge code.

How long does Muse take to reply to a WeChat message?

End-to-end latency is typically 1 to 2 minutes for short tasks. This is governed by WeChat's ~35s long-polling cycle, the ~15s local inbox scan hook, and the worker process launch and execution time.

Why weren't generated images sent to WeChat initially?

Initial bridge scripts usually only implement the text messaging endpoint, saving images to Muse's remote media library. Asking Muse in chat to add image support causes it to implement the WeChat media upload endpoint so images land directly in your chat.

What should I do if the QR code expires?

QR codes expire after 5 minutes. If it times out, simply ask Muse in its web chat to generate a fresh QR code and scan again.

Do WeChat conversations show up in the main Muse web chat?

No. WeChat messaging is handled by an isolated background polling daemon. Chat history stays inside WeChat, while Muse's main interface only logs task execution summaries.

---

Related Guides

• Meta Muse Quick-Start & Waitlist Guide — Account setup, age verification, and starter credits
• Registering Claude Accounts with Meta Muse — Cloud browser automation for overseas accounts
• Automated Cross-Platform Price Tracking with Muse — Hands-on e-commerce scraping and analysis
• Hermes Agent WeChat Integration Guide — Lightweight alternative WeChat AI assistant
• Meta Muse Tutorials Category & All AI Tutorials Hub