Find out what your users say

Both, and they stay separate. AI respondents come with Fiuto and go through the study first, so you get a read before you spend anyone’s time. Real people come two ways: describe the audience you need and Fiuto recruits it from Prolific’s global panel, with the total shown before you order, or share the link with your users, your waitlist, your community. Each run is labelled with which kind answered it, and the two are never mixed. Real people answer anonymously, no account needed.

A persona you give a role to, that goes through your study block by block through the same runner a person uses, answering in its own words. Each one costs 120 credits. It is a fast read, not a stand-in for real people: it shows you where a flow confuses someone before you share the link with anyone.

It reads one surface in your study through a single specialism and reports what it finds. There are four: conversion, whether someone would understand it and act; usability, whether people can get through it; copy, whether the writing could be about any product or only this one; and SEO, whether the page answers what people search for. Which ones you can add depends on the block, since an image carries no links or state to judge. Findings arrive with a severity and sit apart from the responses, so an expert’s opinion never gets counted as someone’s answer.

Point Fiuto at any live URL: a deployed app, a landing page, a working site, or a shared Figma prototype. It records the real path people take through it. If you build with a coding agent (Claude Code, Cursor, Codex, Lovable, v0, Replit, or Bolt), it can draft and launch the study over MCP.

No. The web app does everything: describe what you want to learn, review the plan, launch, read results. The coding-agent path is a shortcut for people already working in one.

Open the link and answer. No sign-up, no install. Screen, mic, or camera recording only starts after they opt in.

Yes, two ways. Point Fiuto at a site in Spanish, German, or another supported language and it drafts the study in that language. Or translate an English study into a new copy in any of 14 languages: every question, option, task, and message. Respondents see the buttons and instructions in their language too. Review the study, then launch it like any other. Translation uses AI credits, on every plan including Free. The help center lists every supported language.

Yes. The free plan launches real studies every month, with every test method and a monthly AI allowance, no card required. Your first study comes with its plan and its analysis on us, and the free allowance covers five AI respondents, so a first run costs you nothing. Raw CSV and JSON export is on every plan too. Exact per-tier limits are in the plans section above.

Eighteen block types, mixable in one study: five-second tests, first-click tests, card sorts, tree tests, preference tests, surveys, hotspot heatmaps, and live-site or prototype tasks.

It drafts the study from a plain description, runs it with AI respondents, inspects your surface as an AI expert, then reads responses back into a takeaway with the quotes behind it. It runs on Anthropic’s Claude. AI features use credits, each plan includes a monthly allowance, and Pro can buy extra packs. No key to paste.