Frequently asked questions

Fiuto is a study authoring and launch tool for testing a prototype or live product with real people. Point it at a URL or a shared prototype, describe what you want to learn, and it drafts a study for you to review before it goes live.

Fiuto is for anyone testing a prototype or product with real users without a dedicated research team: founders, product managers, designers, and teams building with a coding agent. No research background is required.

Yes. Drafting the study, analysing responses into findings, and building the insights deck all work on every tier, including Free, with no AI feature held back for a higher plan.

Same day, in most cases. Describe what you want to learn, review the plan Fiuto drafts, and launch, without a sales call or a procurement cycle.

Yes. Every plan shows its price and what it includes on the pricing section, with no quote request or sales call needed to see a number.

AI credits are what an AI action, like drafting a study or analysing responses, spends. Every plan, including Free, includes a monthly allowance, and Pro can buy extra AI-credit packs on top.

No. Free runs the same Anthropic Claude models Fiuto uses on Pro. Which model handles a task depends on the task itself, not on your subscription tier.

No. Turning your first study's responses into findings costs nothing in credits, once per account, on every tier including Free.

No. Every template in Fiuto's library costs zero credits to start from. Credits are only spent once AI starts drafting or analysing a study.

Eighteen block types you can mix in one study, including five-second tests, first-click tests, card sorts, tree tests, matrix surveys, ranking questions, preference tests, surveys, hotspot heatmaps, and live-site or prototype tasks.

Card sorting asks respondents to group items into categories that make sense to them, useful for building an information architecture from scratch. Tree testing asks them to find a specific item in a structure you've already built, useful for checking one you have.

Yes. Every test method, including card sort, tree test, and matrix survey, is available on every plan, including Free. What stays paid is raw CSV or JSON export, inviting an editor onto a study, and higher launch and response caps; AI is credit-metered the same way on every plan.

Either. A study can send respondents to a working site or a prototype and record the real path they take, so it fits a landing page or checkout flow as easily as an early mockup.

You bring your own. Fiuto doesn't run a paid panel or recruit respondents; you share the study link with people you already reach, such as your users, your waitlist, or your community.

No. Anyone who opens the link answers anonymously, with no sign-up and no install.

No. Fiuto has no per-seat or per-respondent fee. How many responses each study can collect follows your plan's response cap instead.

Only if the study asks for it, and only once the respondent opts in. Recording never starts silently.

In the EU. Study data, recordings, and analytics run on EU-hosted infrastructure.

From Core up. Every result is yours to read and keep in Fiuto on every plan, including Free; raw CSV and JSON export opens up from Core, which keeps a free account from being used to bulk-export data at no cost.

No. Fiuto does not sell personal data or use advertising trackers.

Only people you choose to share it with. Studies and their results are private by default, and publishing a view-only link has no limit on how many people can view it; inviting someone in to edit the study itself is the one thing that needs a paid plan.

It's how a coding agent can draft and launch a Fiuto study directly, without you leaving your editor, over the Model Context Protocol.

No. The web app does everything on its own: describe what you want to learn, review the plan, launch, and read results. The coding-agent path is a shortcut for people already in one, not a requirement.

Claude Code, Cursor, Codex, Lovable, v0, Replit, and Bolt can all draft and launch a study over Fiuto's MCP server, checked 27 July 2026.

Yes. The same MCP connection that drafts and launches a study can also pull its results back into the agent session once responses come in.