XO MCP Server
Connect MCP-compatible AI tools to XO deployments and knowledge-base capabilities.
The XO MCP Server integrates with the XO platform via the Model Context Protocol — giving your AI tools direct control over container deployments and application lifecycle.
Overview
Available tools and technical architecture.
Setup
Install from GitHub or run locally. Configure credentials.
AI Tool Integrations
Connect an MCP client, or use the separate XO GPT flow for ChatGPT knowledge sharing.
Available Tools
| Tool | Description |
|---|---|
deploy_to_xo | Push a containerised app to XO |
start_xo_app | Activate a previously deployed application |
stop_xo_app | Halt a running app without deleting it |
remove_xo_app | Completely remove an application |
get_xo_app_logs | Access real-time application logs |
expose_xo_app | Expose a service with a public URL |
update_knowledgebase_using_text | Update an XO Workspaces knowledge base |
ask_question | Query a knowledge base |
Requirements
- Python 3.11+
uvxpackage manager- Valid XO credentials:
ACCESS_TOKENandDEPLOYMENT_IDfrom XO Launchpad - An MCP-compatible client (Cursor, Claude, Devin, etc.)