Skip to main content
Great Question’s MCP (Model Context Protocol) server gives your team read and write access to studies, sessions, transcripts, highlights, insights, participant data, and study management from Claude Desktop, Claude Code, Cursor, ChatGPT, Copilot or any MCP-compatible AI tool. Agents can also connect entirely on their own, without a person actively driving the session — see Connecting an agent. The MCP covers the full research workflow: candidates, study design and screeners, scheduling, recruitment, session/transcript analysis, and synthesis (highlights, insights, and reels). Full read and write access to your research repository. Agents can search the whole repository at once (sessions, highlights, reels, insights, uploaded files, and survey results ranked together) and follow the links between them: a highlight shows the reels and insights that use it, and an insight carries the highlights and reels it embeds. Agents can also save their work back into Great Question. They can create highlights from a transcript quote, build reels from highlights, write insights with embedded evidence, and create and apply tags. Everything an agent writes is saved under the connected user’s name, follows that user’s existing permissions, and is as findable afterward as anything made in the app.

Setup

Connect Claude, ChatGPT, Cursor, Copilot, or any MCP client.

What's available

Every tool call the server exposes, plus security and data handling.

How to use

Example prompts for working with your research.

Best practices

Get more reliable, efficient results.

Skills

Author reusable instructions your agents follow automatically.
Still need help? Contact us at support@greatquestion.co — median response time is 19 minutes during support hours.