---
title: "Getting started with the Docebo MCP server"
slug: "getting-started-with-the-docebo-mcp-server"
updated: 2026-10-30T13:07:53Z
published: 2026-10-30T13:07:53Z
canonical: "developer.docebo.com/getting-started-with-the-docebo-mcp-server"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://developer.docebo.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Getting started with the Docebo MCP server

## Introduction

The Model Context Protocol (MCP) server allows you to seamlessly integrate Docebo with external AI assistants (MCP clients) like ChatGPT, Claude, Microsoft Copilot Studio, Gemini Enterprise, Slack, and any other compatible MCP clients. Once the integration is configured, users in your organization can interact with Docebo content and data using natural language, directly inside the AI tools they already use, without navigating through the LMS.

This article is intended for Docebo Superadmins and describes how to complete the one-time configuration required to activate the integration.

## How it works

An MCP client sends requests to your Docebo MCP server URL together with a Docebo access token. Each tool runs with the permissions of the Docebo user the token represents, so users can do through the MCP server only what they are already allowed to do in Docebo.

How the token is obtained depends on your scenario:

- External AI assistants (for example, ChatGPT) use OAuth interactive authentication: each user signs in and approves access once.
- Backend services, agents and jobs that you build can use OAuth system-to-system authentication, where no person signs in.

For details, and to choose the best approach for your use case, see [Docebo MCP server authentication and authorization](/v1/docs/docebo-mcp-server-authentication-and-authorization).

## Setting up the Docebo MCP server integration

Setting up the Docebo MCP server integration consists of three steps:

1. Choose and configure the authentication method.
  - For external AI assistants, configure [OAuth interactive authentication](/v1/docs/configuring-oauth-interactive-authentication-for-the-docebo-mcp-server).
  - For backend services, configure [OAuth system-to-system authentication](/v1/docs/configuring-oauth-system-to-system-authentication-for-the-docebo-mcp-server).
2. Configure an MCP server with the tools you want to expose. See [Configuring Docebo MCP servers](/v1/docs/configuring-docebo-mcp-servers).
3. Connect your MCP client. See [Setting up MCP clients for the Docebo MCP server](/v1/docs/setting-up-mcp-clients-for-the-docebo-mcp-server).

Once completed, users can authenticate with Docebo and interact with the tools exposed by the MCP server.

## Available tools

The tools available to users depend on the MCP server they connect to and the tools enabled by the administrator.

The built-in Learner server provides access to learner-focused tools, while custom MCP servers can expose additional capabilities for different personas and use cases.

For a complete list of available tools, including descriptions, permissions, and example prompts, as well as best practices for designing your MCP servers, see the [Supported MCP tools](/v1/docs/supported-mcp-tools) article.