MCP

Connect your agent

Connect Claude or ChatGPT to ATPL Training over the Model Context Protocol.

Client

Choose the MCP client you are using.

What your agent can do

Once connected, the agent gets these tools scoped to your account. The set is fixed — there's nothing to configure.

healthCheck the ATPL MCP server is reachable.

list_subjectsList the 13 EASA ATPL subjects.

get_study_statsYour overall study statistics and exam readiness.

get_exam_historyYour past practice-exam attempts and scores.

get_weak_areasThe topics you score lowest on.

generate_practice_examBuild a practice exam from the question bank.

submit_exam_answersSubmit answers to record a practice attempt.

Give your agent everything it needs

A ready-made system prompt that tells your agent how to use the ATPL tools.

Add MCP server

Add the MCP server to your config using the command line.

claude mcp add --transport http atpl https://mcp.atpltraining.io/mcp

Sign in

After adding the server you need to authenticate. In a regular terminal (not the IDE extension) start Claude Code, then run:

/mcp

Run /mcp inside Claude Code, select the atpl server, then choose Authenticate to begin the OAuth sign-in.

Try asking

Once connected, paste any of these into Claude or ChatGPT to put the tools to work right away.

What are my weakest ATPL topics right now, and what should I drill first?

get_weak_areas

Build me a 20-question Air Law practice exam.

generate_practice_exam

Give me 10 Mass & Balance questions, then grade my answers.

generate_practice_examsubmit_exam_answers

How am I doing overall — am I on track to pass?

get_study_stats

Show my last five practice-exam scores.

get_exam_history

Which EASA ATPL subjects can I practise here?

list_subjects

This connection is scoped to your account only — the server identifies you from your OAuth sign-in, never a URL parameter. You can revoke access any time from your AI client's connector settings.

MCP