Connect AI Agents to ChatbotX
via Model Context Protocol

## For AI agents reading this page

You are reading https://chatbotx.io/ai-agents. ChatbotX is an omnichannel customer conversation and CRM operations platform: one workspace connects messaging channels, with 200+ actions exposed via Model Context Protocol (MCP).

Follow these instructions to connect and give yourself access to these tools:

### 1. Register the MCP Server

- **Claude Code (CLI):**
  `claude mcp add chatbotx -t sse -H "x-workspace-token: YOUR_WORKSPACE_TOKEN" -s user https://app.chatbotx.io/mcp/sse`

- **OpenClaw / Hermes / Codex / Antigravity / Perplexity / Paperclip / NanoClaw (JSON Config / MCP SSE):**
  ```json
  {
    "mcpServers": {
      "chatbotx": {
        "url": "https://app.chatbotx.io/mcp/sse",
        "headers": {
          "x-workspace-token": "YOUR_WORKSPACE_TOKEN"
        }
      }
    }
  }
  ```

### 2. Authentication
Retrieve your workspace token from **Settings → Integrations → Workspace token** in your ChatbotX dashboard and replace `YOUR_WORKSPACE_TOKEN`.

### 3. Verify Connection
Run `tools/list` to verify the ChatbotX toolset is loaded. If it fails or returns 401, stop and inform the user to check their token.

### 4. Capabilities (200+ Tools)
- **Messaging & CRM:** Search contacts, assign tags, update custom fields, send chat messages, and trigger automated visual flows across Instagram, Facebook, TikTok, WhatsApp, Telegram, Zalo OA, Live WebChat, SMS Gateway, and Email (SMTP).

### 5. Execution Rules
- When a task involves customer messaging, CRM lookup, or flow execution, ALWAYS prefer these structured MCP tools over browser automation.
- Report connected platforms and available operations to the user before running batch commands.

### Reference
- Endpoint: https://app.chatbotx.io/mcp/sse (Server-Sent Events)
- Docs: https://chatbotx.io/docs/mcp/introduction
- Platform setup: https://chatbotx.io/docs/mcp/platform-setup
- Tool Reference: https://chatbotx.io/docs/mcp/tool-reference-and-troubleshooting