Artiluca Coach
Practise interviews inside your AI assistant
Artiluca Coach is a free connector that brings real interview practice into Claude, ChatGPT, and other AI assistants: realistic scenarios, and written answers graded on the same published, calibrated rubric used on Artiluca itself. No account needed.
How it works
Ask your assistant for an interview question. It fetches a scenario from Artiluca, you answer in your own words, and it sends your answer back for grading — a 0–100 scorecard with per-dimension comments and a stronger rewrite. Scoring runs on a fixed, calibrated configuration against fixed rubric anchors and is calibration-checked continuously, so revising and re-grading shows real improvement, not model mood. That revise-and-rescore loop is the practice.
Set up
The connector is a remote MCP (Model Context Protocol) server. Its endpoint URL is https://artiluca.ca/api/mcp — no authentication required.
Claude
- Settings → Connectors → Add custom connector (where available on your plan)
- Paste the endpoint URL and add
- Enable it in a chat and ask for an interview question
ChatGPT
- Settings → Apps → Advanced settings → enable Developer mode
- Create an app/connector with the endpoint URL (no authentication)
Other MCP clients — Copilot Studio, Claude Code, Cursor, VS Code, Gemini CLI
Add the endpoint URL as a streamable-HTTP MCP server in your client's MCP settings. Claude Code: claude mcp add --transport http artiluca https://artiluca.ca/api/mcp
Things to try
“Give me an interview question to practise, then grade my answer.”
What happens: The assistant fetches a scenario, waits for your answer, then returns a 0–100 scorecard with six dimension scores and a stronger rewrite.
“I have a job interview this week — run me through the 'Tell me about yourself' opener and score my response.”
What happens: The default interview-opener exercise, then a graded scorecard for your response.
“Grade this answer to the greatest-weakness question, then help me revise it and score the new version.”
What happens: Two scorecards — because scoring is calibrated, the score change between versions reflects the revision.
“Get the presentation-opener exercise and grade my first two sentences.”
What happens: The presentation-opener scenario with tips, then a graded scorecard for your two sentences.
What the connector can do
get_practice_scenarioReturns one of six free exercises — the five interview & early-career scenarios plus the 'Tell me about yourself' opener from the site demo: a realistic scenario, a concrete answering task, and coaching tips. Read-only.
grade_written_answerGrades your written answer on Artiluca's calibrated six-dimension rubric — clarity, structure, concision, precision, audience, impact. Returns a 0–100 overall score, per-dimension comments, strengths, improvements, and a stronger rewrite of your own answer. Read-only.
Both tools are read-only: the connector never modifies anything in your account or your assistant, and it takes no actions on your behalf.
Privacy & limits
- Answers you submit are processed once to generate the scorecard and are never stored — the same no-persistence contract as the site. See the privacy policy.
- The connector needs no account and asks for nothing about you. To enforce fair-use limits we keep only a short-lived (about one day), one-way-hashed caller identifier — the assistant's anonymized per-user id when it supplies one, otherwise a network address — never your answers. Coaching and scores are AI-generated, for practice — not professional advice.
- Grading is subject to fair-use daily limits; answers are capped at 1,200 characters.
- Artiluca's age policy applies: the service is for users 16 or older, or 13–15 with a parent or guardian's permission.
Support
Questions or problems with the connector? Email support@artiluca.ca.