Connect Docebo to Claude: Automate Training Reports and Tracking
Learn how to connect Docebo to Claude using a managed MCP server. This guide covers how to automate LMS reporting, user tracking, and enrollments with AI agents.
If your team uses ChatGPT, check out our guide on /connect-docebo-to-chatgpt-manage-courses-and-learner-enrollments/ or explore our broader architectural overview on /connect-docebo-to-ai-agents-orchestrate-skills-and-certifications/.
If you need to connect Docebo to Claude to automate compliance training reports, manage user enrollments, or track organizational skills, you need a Model Context Protocol (MCP) server. This server acts as the translation layer between Claude's natural language tool calls and Docebo's extensive REST API ecosystem. You can either build and maintain this infrastructure yourself, or use a managed integration platform like Truto to dynamically generate a secure, authenticated MCP server URL.
Giving a Large Language Model (LLM) read and write access to an enterprise Learning Management System (LMS) like Docebo is an engineering challenge. You must handle complex OAuth 2.0 token lifecycles, map massive JSON schemas to MCP tool definitions, and deal with Docebo's strict API quotas and fragmented microservices. Every time Docebo updates an endpoint, you must update your server code, redeploy, and test the integration.
This guide breaks down exactly how to use Truto to generate a secure, managed MCP server for Docebo, connect it natively to Claude Desktop, and execute complex LMS administration workflows using natural language.
The Engineering Reality of the Docebo API
A custom MCP server is a self-hosted integration layer. While the open MCP standard provides a predictable way for models to discover tools, the reality of implementing it against Docebo's APIs is painful. You are not just integrating one "Docebo API" — you are integrating a highly segmented service architecture.
If you decide to build a custom MCP server for Docebo, you own the entire API lifecycle. Here are the specific challenges you will face:
Fragmented Microservice Namespaces
Docebo's API is split across dozens of distinct services: /learn/v1, /manage/v1, /share/v1, /skill/v1, /report/v1, and /audiences/v1, among others. An LLM has no context on which service to query for a specific task. To get a user's enrollment status, do you query /learn/v1/enrollments or /manage/v1/user/activities? You must build an abstraction layer that flattens this namespace and presents a unified set of operations to Claude, hiding the underlying endpoint fragmentation with descriptive tool names.
Asynchronous Reporting and Background Jobs
Not all endpoints in Docebo behave like standard synchronous CRUD operations. When generating heavy training reports or user data extracts (e.g., /report/v1/userreport), Docebo often relies on asynchronous background jobs. The initial API call returns a job_id, and your system must poll another endpoint (/report/v1/userreport/{job_id}) to check the status before the data is ready. If you blindly expose these raw endpoints to Claude, the LLM will hallucinate the polling logic or fail to retrieve the actual data. Your MCP server must intelligently handle these async execution states or provide explicit, chained tool instructions to the model.
Complex Nested Schemas for ILT and Learning Plans
Docebo supports e-learning courses, Instructor-Led Training (ILT), and aggregated Learning Plans. Enrolling a user isn't just toggling a status flag; it requires passing complex JSON payloads that dictate the enrollment level, associated ILT sessions, and optional seat allocations. If an LLM attempts to guess this payload structure, Docebo will return 400 Bad Request errors. A managed MCP server exposes tools like create_a_docebo_enrollment with strictly defined JSON schemas that explicitly guide the LLM to provide the exact required payload structure.
How to Generate the Docebo MCP Server
Truto eliminates the need to manually build tool definitions. By deriving tools directly from Docebo's API documentation and your environment's integration configuration, Truto generates an MCP server dynamically.
Every MCP server is scoped to a single integrated Docebo account and secured by a cryptographic token. You can create this server through the Truto UI or programmatically via the API.
Method 1: Via the Truto UI
For quick testing or one-off administrative agents, the UI is the fastest path:
- Navigate to the Integrated Accounts page in your Truto dashboard and select your connected Docebo instance.
- Click the MCP Servers tab.
- Click Create MCP Server.
- Configure your server (assign a name, select allowed methods like
readorwrite, and set an optional expiration date). - Copy the generated MCP Server URL (e.g.,
https://api.truto.one/mcp/a1b2c3...).
Method 2: Via the Truto API
For production workflows where you deploy AI agents dynamically for your users, you should generate MCP servers via the Truto API. This allows you to securely scope access per tenant.
Make a POST request to the /integrated-account/:id/mcp endpoint:
curl -X POST https://api.truto.one/integrated-account/<DOCEBO_ACCOUNT_ID>/mcp \
-H "Authorization: Bearer <YOUR_TRUTO_API_TOKEN>" \
-H "Content-Type: application/json" \
-d '{
"name": "Docebo Compliance Auditor Agent",
"config": {
"methods": ["read"],
"tags": ["reports", "users"]
},
"expires_at": "2026-12-31T23:59:59Z"
}'The response returns your managed MCP server URL:
{
"id": "mcp_srv_98765",
"name": "Docebo Compliance Auditor Agent",
"url": "https://api.truto.one/mcp/a1b2c3d4e5f6..."
}This URL is fully self-contained. It handles all Docebo OAuth token refreshes, pagination translation, and schema validation automatically.
Connecting the MCP Server to Claude
Once you have your Truto MCP URL, you can plug it into Claude. Because Truto MCP servers use the standardized Server-Sent Events (SSE) JSON-RPC protocol over HTTP, no extra middleware is required.
Method A: Via the Claude Desktop UI (or ChatGPT)
If you are using ChatGPT or Claude Desktop's visual interface:
- Open Settings -> Integrations (in Claude) or Connectors (in ChatGPT).
- Click Add MCP Server (or Custom Connector).
- Name the connector (e.g., "Docebo LMS").
- Paste the Truto MCP Server URL.
- Click Add / Save.
The model will immediately handshake with the Truto endpoint and load the available Docebo tools.
Method B: Via the Claude Config File
If you are configuring Claude Desktop manually or running a headless LangChain/LangGraph agent, you can define the SSE connection in your configuration file.
Add the following to your claude_desktop_config.json:
{
"mcpServers": {
"docebo_lms": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-sse",
"https://api.truto.one/mcp/a1b2c3d4e5f6..."
]
}
}
}Restart Claude Desktop, and the "docebo_lms" toolset will be available in your prompts.
Docebo Hero Tools for AI Agents
Truto exposes hundreds of Docebo endpoints, but for training reports, enrollment automation, and skills tracking, a few specific tools provide the highest leverage for LLMs.
Here are the hero tools your agent will use most often:
list_all_docebo_courses
Retrieves a paginated list of all active courses in the Docebo catalog. Essential for agents that need to cross-reference course IDs before executing enrollments.
"Fetch a list of all active courses in our Docebo catalog that relate to 'Cybersecurity' and output their course IDs and titles."
docebo_course_enrollments_list_2
Retrieves course enrollments across the LMS. This is the primary tool for auditing who is currently taking a course and checking their completion status.
"Pull the enrollment list for the course ID 4092. Filter the results to show me only users who have a status of 'in_progress' and have been enrolled for more than 30 days."
create_a_docebo_enrollment
Enrolls one or more users into a specific course. The tool schema enforces Docebo's requirement for user IDs, course IDs, and the enrollment level (e.g., learner vs tutor).
"Enroll the user with ID 8849 into the 'Managerial Ethics 2026' course (ID 5512). Set their enrollment level to learner."
docebo_user_get_activities
Fetches detailed activity information for a specific user, including their visited sessions per month, time spent, and enrollment velocity. Perfect for deep-dive performance reviews.
"Get the detailed learning activities and session history for user ID 1029 over the last 90 days. Summarize their total time spent in courses."
docebo_report_report_get_data
Extracts row data from a pre-configured Custom Report in Docebo (using the report ID). This is vital for extracting massive compliance matrices without hitting individual user endpoints.
"Extract the row data from custom report ID 15 (Annual Compliance Matrix). Cross-reference the output to find any employees in the Engineering branch who are marked as non-compliant."
docebo_manager_subordinates_get_candidates
Lists all subordinates in the organizational chart for a specific manager. AI agents use this tool to determine the blast radius of team-specific enrollment workflows.
"List all the subordinates for manager ID 441. Format the response as a table with their names, user IDs, and current branch."
For the complete inventory of available Docebo endpoints and their exact JSON schema requirements, visit the Docebo integration page.
Workflows in Action
With these tools available, Claude can execute complex LMS operations autonomously. Here are two real-world workflows you can run using the Truto MCP server.
Workflow 1: The Team Compliance Audit
A department head wants to know which of their team members have fallen behind on mandatory security training. Instead of navigating Docebo's complex reporting UI, they simply ask Claude.
"I need an audit of my team's progress on the 'Q3 Security Awareness' course (Course ID 104). Find all my subordinates, check their enrollment status for that course, and give me a list of anyone who hasn't completed it yet."
Execution Steps:
- Claude calls
docebo_manager_subordinates_get_candidates(passing the manager's context ID) to retrieve the list of subordinate user IDs. - Claude calls
docebo_course_enrollments_list_2for Course ID 104. - Claude correlates the two datasets in memory, matching the subordinate IDs against the enrollment completion statuses.
- Claude outputs a formatted list of non-compliant employees directly in the chat.
sequenceDiagram participant Claude as Claude participant MCP as Truto MCP Server participant Docebo as Docebo API Claude->>MCP: Call docebo_manager_subordinates_get_candidates MCP->>Docebo: GET /manage/v1/managers/subordinates/candidates Docebo-->>MCP: Returns team list array MCP-->>Claude: JSON response Claude->>MCP: Call docebo_course_enrollments_list_2 MCP->>Docebo: GET /course/v1/courses/enrollments Docebo-->>MCP: Returns enrollment statuses MCP-->>Claude: JSON response
Workflow 2: Bulk Onboarding Enrollment
HR operations often require enrolling new cohorts into standardized Learning Plans (which group multiple courses together).
"We have a new cohort starting today. Retrieve the IDs for users: jsmith, adoe, and rroe. Then, enroll all of them into the 'Global Engineering Onboarding' Learning Plan (Plan ID 88) as learners."
Execution Steps:
- Claude calls
list_all_docebo_user(with a search query) to look up the internal Docebo user IDs for jsmith, adoe, and rroe. - Claude structures the bulk payload required by Docebo's schema.
- Claude calls
docebo_learningplan_enrollments_create_bulk, passing Plan ID 88 and the array of resolved user IDs. - Claude confirms the successful enrollments to the operator.
Security and Access Control
Exposing an enterprise LMS to an AI agent requires strict guardrails. Truto's MCP architecture provides multiple layers of security at the token level, ensuring your agents only access what they strictly need.
- Method Filtering: By defining
config.methods: ["read"]during server creation, you can physically prevent the LLM from executing destructive actions (like deleting courses or un-enrolling users). The server will simply omit allPOST,PUT, andDELETEtools from the model's context. - Tag Filtering: You can restrict the MCP server to specific API domains. Passing
config.tags: ["report", "user"]ensures the LLM only sees tools related to reporting and user directories, hiding unrelated administrative tools like billing or integrations. - API Token Authentication: By enabling
require_api_token_auth: true, possession of the MCP URL is no longer enough to invoke tools. The client (or agent framework) must also pass a valid Truto API token in theAuthorizationheader, enforcing identity validation. - Time-To-Live (TTL): The
expires_atconfiguration creates ephemeral MCP servers. Once the timestamp passes, Truto's infrastructure automatically cleans up the server and destroys the token, making it ideal for temporary contractor access or limited-time audit agents.
Handling Docebo Rate Limits
When writing automated workflows, rate limits are a critical architectural consideration. Docebo enforces strict API quotas based on your subscription tier.
Factual note on rate limits: Truto does not retry, throttle, or apply backoff on rate limit errors. When the upstream Docebo API returns an HTTP 429 (Too Many Requests), Truto passes that error directly back to the caller.
However, Truto normalizes Docebo's upstream rate limit information into standardized headers (ratelimit-limit, ratelimit-remaining, ratelimit-reset) per the IETF specification. This makes it much easier for your AI agent framework (like LangGraph or CrewAI) to catch the 429 error, read the standardized ratelimit-reset header, and execute its own intelligent backoff and retry logic. You maintain complete control over how your agent handles capacity constraints.
Wrapping Up
Connecting Claude to Docebo transforms how organizations manage compliance, skills tracking, and learning administration. Instead of forcing managers and HR teams to navigate complex LMS dashboards, you can provide an agentic interface that executes complex reporting and enrollment workflows via natural language.
By leveraging Truto's managed MCP servers, you bypass the massive engineering overhead of reading documentation, writing custom OpenAPI tool wrappers, and managing OAuth lifecycles. Truto handles the protocol translation and authentication, letting your engineering team focus entirely on prompt engineering and workflow orchestration.
Ready to give your AI agents secure access to Docebo? Get started with Truto today.
FAQ
- Can I restrict Claude to read-only access for Docebo?
- Yes. When creating the Truto MCP server, you can set the configuration to only allow 'read' methods. This guarantees the LLM can query training reports and user activities, but cannot enroll users, delete courses, or alter LMS configurations.
- How does Truto handle Docebo's background report jobs?
- Truto exposes the underlying Docebo endpoints for background reports. The LLM can use the initial endpoint to trigger a report and receive a job ID, and then use the status endpoint to poll for completion, replicating the native API flow.
- Do I need to manage Docebo OAuth tokens manually?
- No. The Truto MCP server abstracts the entire authentication lifecycle. You provide the unified Truto MCP URL to Claude, and Truto automatically handles connecting to Docebo and refreshing the OAuth tokens securely.
- How are Docebo rate limits handled in the MCP server?
- Truto does not absorb or auto-retry rate limits. If Docebo returns an HTTP 429 error, Truto passes it directly to Claude along with standardized IETF rate limit headers. Your client or agent framework is responsible for implementing retry/backoff logic.