Your first learning session

Start with two working connections

In Tutor ACP, connect your chosen agent and your Tutor service. The agent provides the conversation; Tutor keeps the course and learning state. Your institution supplies its Tutor address and authorises your membership.

If you have not connected yet, follow Install Tutor ACP. If you use another AI application, follow Connect an MCP client instead.

A message written before setup is complete stays saved as Ready to send. Finish connecting in the same window to continue.

Choose a course or bring your own goal

Open Courses to find your personal programmes and the shared courses available in your institution. Read the programme, check its access conditions and choose the course you want to follow. An invitation or approval may be required.

If your institution allows personal courses, start in Conversation with a subject you care about. Tutor looks for an existing programme before creating a new private one, and saves a new programme from the first exchange.

I want to learn Go so I can build a small web API.
I already know Python and can practise for 20 minutes a day.
Help me find my starting point and guide me through a first session.

In a general MCP client, add “Use Tutor MCP” to make clear which learning service the assistant should use.

Try an activity, then explain your thinking

The tutor can check what you already know, explain a concept and propose an exercise. Answer in your own words and show your reasoning. If you needed a hint, say so: unaided recall and assisted practice provide different evidence.

You can change the pace without abandoning the course. Ask for a smaller example, a different explanation, a review or a challenge closer to your own project.

Explain this with a smaller example.
Then give me one exercise without showing the answer.
After I try it, help me understand any mistakes.

Check any requested tool permission before allowing it. In Settings → Agents, permissions can be remembered for your selected account, Tutor service and agent, then changed later.

Understand what your progress means

Open the course and choose Progress. Tutor keeps review dates and learning evidence so the next session can build on previous work.

• Estimated: a model uses your recorded responses to estimate what you know.
• Retained: eligible delayed recall provides evidence that learning lasted.
• Demonstrated: an eligible assessment supports a skill claim.
• Transferred: eligible assessments show use across different situations.

A high estimate alone does not prove mastery, and badges are not an external qualification. Treat missing evidence as something to practise, rather than a score to maximise.

Finish with a next step

Summarise what we worked on, including what I still find difficult.
Save the session in Tutor and suggest a useful next step.
Tell me what is due for review.

Next time, select the same account and Tutor service, reopen the conversation or choose the course in Courses, and ask to continue. Each conversation has its own active course; select another course explicitly when you want to switch.

Your server keeps the learning record, while Tutor ACP keeps conversation files on your device. See Backups and updates to protect both.

Keep going.