v1.2.0 Feature

Run UX research in Claude via MCP

Zernote MCP lets UX researchers run a research loop in Claude: list folders, open interviews, read notes first, search for evidence, and cite quotes with timestamps.

Claude explaining the Zernote research loop: upload, ask questions, structure emerges
In Claude: upload, ask in plain language, build structure from your corpus

What

Zernote in Claude is a research library. Work in order: folders, then interviews, then notes and summary.

Search or cards supply evidence. Open a transcript window when you need more context.

Writing notes back requires your explicit ask. We append by default and only replace when you say so.

Why

Loading full transcripts into the model burns tokens and hides the useful lines. Notes compress an interview once so you return without reloading the whole call.

The server stays thin. The research loop runs in the agent. Zernote stores evidence. The agent handles conversation.

How to use

  • Connect Zernote in Claude.
  • List folders. Pick a project. List interviews.
  • Open an interview. Read notes, summary, JTBD first.
  • Need evidence? Search or cards. Use quote links over paraphrase.
  • Still thin? Pull a short transcript window around the relevant chunk.
  • Found something worth keeping? Say so and we append it to notes.

Good to know

Next session: read notes again. Skip loading the full transcript into context first.

Ask explicitly if you want a note replaced. Default behavior is append.

Topics