| Reos | Maze | |
|---|---|---|
| Plan | Yes | Partial |
| Recruit | Yes | Yes |
| Conduct | Yes | Yes |
| Ingest | Yes | Partial |
| Analyze | Yes | Partial |
| Act & close the loop | Yes | Partial |
Reos vs Maze
Maze shows you where. Reos also asks why.
Maze is a strong unmoderated-testing platform with a large panel; its interview studies and AI moderator sit on Enterprise. Reos runs the tests too, puts AI-moderated and live interviews on every plan, and analyzes them next to your tickets and calls.
The honest take
Testing at scale - or the whole conversation
If the job is fast unmoderated testing with a big panel, Maze is very good at it. The difference is how far you can go once a test shows you where people get stuck.
Choose Maze if...
- Unmoderated testing is most of your research: prototype, first-click, card sort, tree test and live website tests, every week.
- You need a big panel fast - millions of participants in 150+ countries, with incentives included in the credits.
- Your designers work in Figma and tools like Figma Make, Lovable and Bolt, and want testing wired straight into them.
Choose Reos if...
- You want the why as well as the where: AI-moderated and live interviews on every plan, not only on Enterprise.
- You want test results, interviews, tickets and calls analyzed together, with each claim checked against its quote.
- You want a published price before you talk to anyone, and your research data hosted in the EU if you need it.
Side by side
The loop, side by side
Maze goes deep on unmoderated testing and its panel. Here is how the rest of the research workflow compares.
| Feature | Reos | Maze |
|---|---|---|
| Plan | ||
| AI study-planning assistantplan, research questions & guide from a conversation | Yes | Partial, screeners, up to 10 questions |
| Recruit | ||
| Your own participantsCSV lists, segments & in-product forms, consent on file | Yes, every plan | Partial, full Reach database on Enterprise |
| A large external panel | Partial, Prolific, enabled on request | Yes, millions of participants, 150+ countries |
| Incentives handled | Partial, your own Tremendous account, on Enterprise | Yes, included in panel credits |
| Scheduling for live sessionswith reminders | Yes | Partial, Enterprise |
| One open link, no recruiting step | Yes | Yes |
| Conduct | ||
| AI-moderated interviewstwo-way voice, sessions in parallel, listen in live | Yes, every plan, 2 units each | Partial, Enterprise only |
| Interview languages | Yes, 12 languages | Yes, 20 languages listed |
| Live interviews, hosted in the browserrecorded & transcribed, silent observers | Yes | Partial, Enterprise only |
| Unmoderated testsfirst-click, five-second, card sort, tree test | Yes, closed card sorts | Yes, also open card sorts |
| Figma prototype testsmisclicks & rage clicks tracked | Yes, every plan | Yes, with a first-click task |
| Live-website tests | Yes, every plan | Yes |
| Recorded think-aloud tasksscreen, mic & camera | Yes, every plan | Partial, clips, on paid plans |
| Mobile app tests | Not offered | Yes, Enterprise |
| Meeting bot for Meet, Zoom & Teams | Yes, 1 unit per meeting | Not offered |
| Ingest | ||
| Recording uploads & transcription | Yes | Yes, up to 1 GB and 8 hours |
| Ticket, email & call connectorsIntercom, Front, Gong | Yes | Not offered, none among its integrations |
| Analyze | ||
| AI themes & reports | Yes | Partial, thematic analysis on Enterprise |
| Every claim checked against its quotethe DIVE™ pipeline drops what the quote doesn’t support | Yes | Partial, traceable quotes, no claim check described |
| Cross-study Super Insights | Yes | Partial, across studies only through MCP |
| Act | ||
| Initiatives synced with Linear & Jirastatus flows back | Yes | Partial, Jira integration |
| MCP for AI assistantsClaude, ChatGPT, Gemini & others | Yes, read & write, scoped by role | Partial, read-only, Enterprise |
| Pricing & trust | ||
| Interviews without a sales callpublished plans from $59 a month | Yes | Not offered, Enterprise is priced on request |
| A free plan | Not offered | Yes, 5 seats, 1 study a month |
| Research data hosted in the EU | Yes, Frankfurt or the US | Not offered, AWS us-east-1 |
| SOC 2 Type II & ISO 27001 | Yes | Yes |
| ISO 42001 for AI management | Yes | Not offered, not found on its pages |
based on Maze’s public documentation, October 2026; Reos column checked October 2026 · spotted something out of date? tell us - we’ll fix it
Where and why
A click map can’t tell you what people meant
A first-click test shows you where people get stuck, at scale - and Maze does that well. The reason sits in what people say: in an interview, a support ticket, a recorded call. Reos keeps the test and the conversations in one project and analyzes them together.
- Tests on every planFigma prototype tests with misclicks and rage clicks, live-website tests, recorded think-aloud tasks, first-click, five-second, preference and navigation tests, closed card sorts and tree tests.
- Then ask whyFollow a test with AI-moderated interviews the same day, or run them yourself, hosted in the browser - both on every plan.
- One analysis across it allTest results, interviews, tickets and calls land in the same project, and a claim survives only if its quote supports it.
an example study · where, then why
where · first-click test
most first clicks miss the export button
why · in their words
- AI interview“The grey text looked like a label, not a button.”
- Support ticket“Where do I export the report? I can’t find it.”
- Live interview“For anything like that I go to Settings first.”
insight
The export button reads as a label, so people look for it in Settings.
4 sources · each claim checked against its quote
an illustration - your own studies fill it in
Interviews
Interviews from the first plan
On Maze, interview studies and the AI moderator are Enterprise features. On Reos they come with every plan, priced in units you can see: an AI-moderated interview is 2 units, a hosted, uploaded or bot-recorded one is 1.
- AI-moderated, around the clockTwo-way voice in 12 languages and sessions in parallel; listen in on any of them live. Participants are told they are talking to an AI.
- Or run it yourselfHosted video interviews with the participant’s screen in view, silent observers, scheduling with reminders, and feedback on your moderation afterwards.
- Meetings you already haveSend the Reos bot to Meet, Zoom or Teams for 1 unit, or upload a recording from any tool - transcribed, with speakers detected.
where interviews start
Maze plans
- Free$0Tests and surveys, 1 study a monthNo interviews
- Starter$99 a month, where offeredThe unmoderated toolkit and recordingsNo interviews
- EnterprisePriced on requestInterview studies, AI moderator, thematic analysisInterviews included
Reos plans
- Researcher$59 a month12 units a month, 3 seatsInterviews included
- Team$249 a month60 units a month, 15 seatsInterviews included
- Scale$999 a month300 units a month, 40 seatsInterviews included
monthly prices as published, October 2026 · Reos annual plans are $49, $208 and $833 a month
Pricing
Interviews on the plan, not the contract
Maze’s free plan is a generous way to start testing. Interviews, the AI moderator and thematic analysis come with Enterprise, priced on request. Reos publishes every plan, from $59 a month, with research units included.
Maze
- A real free planFive seats and one study a month at no cost - a genuine way to start testing.
- Interviews on EnterpriseInterview studies, the AI moderator and thematic analysis come with the Enterprise plan, priced on request.
- Panel creditsPanel responses are paid in credits with incentives included: 3 credits per unmoderated response, pay as you go on Free and Starter.
Reos
- One monthly priceHosting, your team, panel and repository - from $59 a month, with research units included.
- Units, not metersAn interview or a written report is 1 unit, an AI-moderated interview 2, and 100 tickets or reviews 1. Transcription, analysis, clustering and storage come with it.
- Top up, never guessHeavier month than planned? Top up units in one click at a published rate, and they never expire.
Switching
Bring your research with you
Bring your session recordings and result exports into Reos - uploads are transcribed and time-coded, CSV exports become observations, and your next study starts from the evidence you already have.
Recordings & transcripts
Upload sessions from any tool - transcription, speakers, and time-codes are automatic.
Datasets of any size
Survey exports, review dumps, call logs - CSV or JSON, mapped into analyzable observations.
Channels, connected
Point your ticket queue, inbox, and forms at Reos once - new data flows in from then on.
Run your next study in Reos
Sign up, import an interview, and trace your first insight to its source in minutes.