Monotree MCP server
The MCP server lets AI assistants work with the content and people in your Monotree platform. Once connected, an assistant can answer questions from your handbooks, look up people and departments, pull training and engagement figures, and draft new content for you to review.
MCP (Model Context Protocol) is an open standard that AI clients such as Claude, ChatGPT and Gemini use to talk to external systems. Monotree exposes one MCP server per company:
https://customer.monotree.com/mcpWhat you can do
- Get reports. Onboarding and course completion, engagement, checklist compliance, expiring certificates, survey results and form responses.
- Draft content. New handbooks, courses, chapters, quizzes, announcements, events and journeys are created as drafts. A person reviews and publishes them in the CMS.
- Ask about content. "What does our handbook say about sick leave?" The assistant searches handbooks, courses, announcements and events and reads the matching chapter.
- Look up people. Find employees, see who manages a department, and walk the org chart.
Who can use it
People sign in with their Monotree account. Admins, editors and managers add the server URL to Claude, ChatGPT or Gemini, click Connect and sign in. The assistant then sees exactly what that person can see in the CMS. Other employees cannot connect.
Automations use an API token. For coding agents, scripts and custom agents, an admin creates a token in the CMS and adds it to the client. The token acts as that admin and sees everything its scopes allow.
Both are described on the Authorization page.
Where to go next
- Getting started connects your first assistant in three steps.
- Connect a client has instructions per client.
- Working with an assistant explains drafts, confirmations and good prompts.
- Tools reference lists every tool the assistant can call.