Honne
Honne runs AI-moderated user research: prototype and live-website tests, five-second tests, surveys and interviews, where an AI moderator asks each participant the follow-up questions a researcher would. With the Honne connector, Claude works with your studies directly: - Read results: session totals, each participant's answers and moderated transcript, AI-extracted insight themes with quotes, and the latest research report. - Build studies: create a study, add sections with participant tasks and settings, write the AI moderator's brief, then publish it and get the participant link. Publishing runs the same checks as the Honne app. - Run persona panels: research your customers once to build your workspace's synthetic personas, then run them through any study's design or compare two designs to get a ranked list of what to fix next. Claude acts with your own Honne permissions and only sees the workspaces you belong to. Every change asks for your confirmation, and Claude can't delete studies, sections, personas or designs. Participant answers are labelled as untrusted research data. Disconnect any time under Settings, Profile, Connected apps in Honne.
Details
https://tryhonne.com/api/mcp · streamable-http
