OpenClaw WeCom Plugin (Bone Version - Multi-App Support)

This is a basic, locally developed OpenClaw plugin for Enterprise WeChat (WeCom). It supports concurrent integration with multiple WeCom applications, including both Intelligent Robot and Self-Built Apps. It handles core message parsing, basic AI integration, and appropriate responses (streaming JSON or passive XML).

Key Features

• Concurrent Multi-App Support: Integrate multiple WeCom Intelligent Bots and Self-Built Apps simultaneously.
• Dual Mode Per-App: Each configured application can operate as either an Intelligent Bot (JSON streaming) or a Self-Built App (XML passive replies).
• Message Parsing: Decrypts and parses incoming messages in both JSON (AI Bot) and XML (Self-Built App) formats, dynamically detected per application.
• Basic AI Integration: Routes parsed messages to OpenClaw's AI core for processing.
• Flexible Replies: Generates streaming JSON responses for AI Bots or passive XML replies for Self-Built Apps.
• Security: Utilizes WeCom's message encryption/decryption, dynamically validates the CorpID from inbound messages for Self-Built Apps, for secure communication.
• Unique Webhook Paths: Each application can be configured with its own dedicated webhook URL for clear separation.

Prerequisites

• OpenClaw installed (version 2026.1.30+)
• Enterprise WeChat admin access to configure Intelligent Robot or Self-Built applications.
• A public-facing server address accessible from Enterprise WeChat (HTTP/HTTPS) for webhooks, with unique paths for each application.

Installation

To install this plugin locally:

1. Copy the plugin directory to your OpenClaw extensions folder:

    cp -r /mnt/codeserver/project/WeComCompanyPlugin ~/.openclaw/extensions/

2. Enable the plugin and configure applications in openclaw.json:

   Add or modify the channels.wecom section in your OpenClaw configuration file (~/.openclaw/openclaw.json). You will define an array of applications, where each object configures a distinct WeCom integration.

    {
      "plugins": {
        "entries": {
          "WeComCompanyPlugin": {  // This key MUST match your plugin folder name
            "enabled": true
          }
          // ... other plugins if any
        }
      },
      "channels": {
        "wecom": {
          "enabled": true,
          "applications": [
            {
              "id": "my_ai_bot", // Unique ID for your AI Bot application
              "enabled": true,
              "token": "YOUR_AI_BOT_TOKEN",
              "encodingAesKey": "YOUR_AI_BOT_ENCODING_AES_KEY",
              "isSelfBuiltApp": false, // Set to false for AI Bot
              "webhookPath": "/webhooks/wecom/ai_bot" // Unique URL path for this bot
            },
            {
              "id": "my_self_built_app", // Unique ID for your Self-Built App application
              "enabled": true,
              "token": "YOUR_SELF_BUILT_TOKEN",
              "encodingAesKey": "YOUR_SELF_BUILT_ENCODING_AES_KEY",
              "isSelfBuiltApp": true, // Set to true for Self-Built App
              "webhookPath": "/webhooks/wecom/self_built" // Unique URL path for this app
            }
          ]
        }
      }
    }

   Replace placeholder values with your actual WeCom application details. Ensure webhookPath is unique for each application. If webhookPath is omitted, it defaults to /webhooks/wecom/<application_id>.

3. Restart OpenClaw for the changes to take effect.

    openclaw gateway restart

Configuration Options

Under channels.wecom.applications (an array of objects):

Enterprise WeChat Configuration

Follow these steps in the Enterprise WeChat Admin Console for each application you configure:

For Intelligent Robot (AI Bot) Integrations

1. Log in to the Admin Console.
2. Navigate to "Application Management" > "Applications" > "Create Application" > Select "Intelligent Robot".
3. Configure "Receive Messages":
   • URL: Use the webhookPath configured for this AI Bot in openclaw.json (e.g., https://your-domain.com/webhooks/wecom/ai_bot). Ensure this URL is publicly accessible.
   • Token: Match the token specified for this AI Bot in openclaw.json.
   • EncodingAESKey: Match the encodingAesKey specified for this AI Bot in openclaw.json.
4. Save and enable message receiving.

For Self-Built App Integrations

1. Log in to the Admin Console.
2. Navigate to "Application Management" > "Applications" > Select your existing Self-Built App or create a new one.
3. Configure "Receive Messages":
   • URL: Use the webhookPath configured for this Self-Built App in openclaw.json (e.g., https://your-domain.com/webhooks/wecom/self_built). Ensure this URL is publicly accessible.
   • Token: Match the token specified for this Self-Built App in openclaw.json.
   • EncodingAESKey: Match the encodingAesKey specified for this Self-Built App in openclaw.json.
4. Save and enable message receiving.

Important: For Self-Built Apps, ensure isSelfBuiltApp is set to true in your openclaw.json configuration for that specific application.

FAQ

Q: How does inbound message decryption and parsing work for multiple applications (AI Bot JSON and Self-Built App XML)?

A: The plugin dynamically identifies the target application based on the incoming webhook URL. For each request, it:

1. Uses the token, encodingAesKey, and optionally corpId associated with the identified application.
2. Automatically detects if the incoming message body is JSON (for AI Bots) or XML (for Self-Built Apps).
3. Extracts and decrypts the message content from the respective format.
4. For Self-Built Apps, it extracts the CorpID from the decrypted inbound message and uses it for constructing secure passive replies.
5. Parses the decrypted content (JSON or XML) into a normalized message object for consistent processing by the AI.

Q: How does outbound image sending work?

A: The plugin automatically handles images generated by OpenClaw (such as browser screenshots):

• For AI Bot (streaming JSON): Local images (from ~/.openclaw/media/) are automatically encoded to base64 and sent via WeCom's msg_item API. Images appear when the AI completes its response.
• For Self-Built App (passive XML): Direct passive XML replies do not support sending images from local paths. If the AI generates an image, its local path will be converted into a text description within the passive XML text reply.

Project Structure

WeComCompanyPlugin/
├── index.js              # Plugin entry point (main plugin logic and HTTP handler, multi-app routing)
├── wecom-message-processor.js # Core logic for message processing and dispatching
├── webhook.js            # WeCom HTTP communication handler, message/event parser (JSON/XML), and reply builder
├── crypto.js             # WeCom encryption algorithms (message + media)
├── logger.js             # Logging module
├── stream-manager.js     # Streaming response manager (for AI Bot mode)
├── utils.js              # Utility functions (TTL cache, deduplication)
├── package.json          # npm package config
└── openclaw.plugin.json  # OpenClaw plugin manifest