Omnichannel MCP Server for OpenAI Codex

# ChatbotX Model Context Protocol (MCP) Server for OpenAI Codex

Connect OpenAI Codex developer agent to ChatbotX via MCP. Automate customer conversations, CRM contact resolution, webhook inspections, and chatbot flow pipelines directly from code and developer runtime environments.

## Connection Setup for OpenAI Codex
- Protocol: Model Context Protocol (MCP) via Server-Sent Events (SSE)
- Endpoint: https://app.chatbotx.io/mcp/sse
- Authentication: Header "x-workspace-token: <YOUR_WORKSPACE_TOKEN>"
- Documentation: https://chatbotx.io/docs/mcp/introduction

### Configuration (~/.codex/config.toml)
```toml
[mcpServers.chatbotx]
url = "https://app.chatbotx.io/mcp/sse"

[mcpServers.chatbotx.headers]
x-workspace-token = "YOUR_WORKSPACE_TOKEN"
```

## Available MCP Tools
- contacts_search: Programmatically lookup contacts by email, phone, or workspace ID.
- contacts_tag: Add or remove segmentation tags during code execution loops.
- contacts_custom_field: Mutate or query custom attributes on contact objects.
- flows_trigger: Trigger automated chatbot sequences and webhook handlers.
- conversations_list: Query active conversation threads across messaging providers.
- messages_send: Dispatch outbound API messages and transactional notifications.
- broadcasts_list: Inspect broadcast campaigns and delivery analytics.
- channels_list: Verify connected channels and active webhook listener health.

## Developer Workflows with Codex
```bash
$ codex "Check webhook delivery status for contact +84901234567 on WhatsApp and trigger flow_order_status"
$ codex "Query CRM for contacts with custom_field:plan='enterprise' and verify their Telegram handle"
$ codex "Automate test broadcast payload dispatch and verify SSE listener response"
```

## Supported Messaging Channels
WhatsApp, Messenger, Instagram, TikTok, Telegram, Zalo OA, Webchat, SMTP, and API.