openclaw-plugin-wecom  by sunnoy

Enterprise WeChat AI integration plugin

Created 1 month ago
472 stars

Top 64.7% on SourcePulse

GitHubView on GitHub
Project Summary

This plugin integrates Enterprise WeChat (WeCom) AI capabilities into the OpenClaw framework, enabling users to leverage advanced AI bots within their WeCom environment. It targets OpenClaw users seeking seamless AI integration, offering robust features like multi-mode support, reliable message delivery, dynamic agent management, and multimedia handling, thereby enhancing communication and automation within WeCom.

How It Works

The plugin supports three primary message modes: AI Bot Mode (JSON, streaming output), Agent Mode (XML, full message capabilities), and Webhook Bot Mode (for group notifications). Its core design features a four-layer message delivery fallback mechanism (streaming, response_url, Webhook Bot, Agent API) to ensure message reliability. Dynamic Agent Management automatically provisions isolated agents per user or group chat, maintaining distinct conversational contexts. It also incorporates message debouncing to consolidate rapid user inputs into single AI requests and handles various multimedia types, including encrypted image decryption and file uploads/downloads.

Quick Start & Requirements

  • Primary install: bash\nopenclaw plugins install @sunnoy/wecom
  • Prerequisites:
    • OpenClaw (version 2026.1.30+)
    • Enterprise WeChat (WeCom) admin backend access for application creation.
    • A server address accessible from WeCom (HTTP/HTTPS).
  • Links:
    • WeCom AI Bot Docs: https://work.weixin.qq.com/wiki/doc/api?level=90&tag=90401
    • WeCom Self-Built App Docs: https://work.weixin.qq.com/wiki/doc/api?level=90&tag=90402
    • WeCom Group Bot Docs: https://work.weixin.qq.com/wiki/doc/api?level=90&tag=90403

Highlighted Details

  • Multi-layer Message Delivery Fallback: Ensures message reliability through streaming, response_url, Webhook Bot, and Agent API layers.
  • Dynamic Agent Management: Automatically creates isolated agents per user/group chat for strong data isolation and context management.
  • Multimedia Support: Handles text, images (including decryption of incoming encrypted images), voice (auto-transcribed in private chats), and files.
  • Message Debouncing: Consolidates rapid user messages into a single AI request to prevent redundant LLM calls.
  • Webhook Bot Mode: Facilitates sending notifications to group chats via configured Webhooks.

Maintenance & Community

Contribution guidelines are available in CONTRIBUTING.md. Specific details on maintainers, sponsorships, or community channels (like Discord/Slack) are not explicitly provided in the README.

Licensing & Compatibility

  • License: ISC License.
  • Compatibility: Permissive license suitable for most integration scenarios. Requires OpenClaw 2026.1.30+ and specific WeCom backend configurations.

Limitations & Caveats

The AI Bot Mode strictly requires JSON callbacks, necessitating API mode creation in WeCom; standard mode (XML) is unsupported. Agent Mode supports XML callbacks. Agent API calls may be blocked (error 60020) if the server's outbound IP is not whitelisted in WeCom's trusted IP settings. Outbound image sending has limits (2MB/image, 10 images/message). Voice message transcription is limited to private chats.

Health Check
Last Commit

1 day ago

Responsiveness

Inactive

Pull Requests (30d)
9
Issues (30d)
69
Star History
346 stars in the last 30 days

Explore Similar Projects

Feedback? Help us improve.