Maze

1. Overview

Maze (maze.co) is a user research platform. It covers unmoderated prototype and usability testing, surveys, card sorting and tree testing, live website and mobile testing, moderated interview studies, an AI moderator, and a paid participant panel (Pricing, Homepage). For this section, the question is narrow: can one product builder check a problem and their success criteria with real users, with no research team? Mostly yes, for unmoderated, task-based work. The conversational, discovery-grade part (AI moderator) sits behind Enterprise.

Surfaces to keep distinct

Surface	What it is	Plan signal on public pricing page	Sources
Prototype testing	Import a Figma or Axure prototype, or AI-generated prototypes from Figma Make, Lovable, Bolt and Replit. Missions, first-click tasks, up to five variants per study, heatmaps, misclick and completion metrics, plus a vendor “Usability Score”	All plans	Prototype testing
Surveys, card sorting, IA tests	Unmoderated question and IA blocks	Surveys and usability testing: all plans. Live website and mobile testing: Enterprise	Pricing, Surveys
Panel	Pay-per-response recruitment. Most orders go through a Core Panel “powered by Prolific”, with an Extended Network for other regions and B2B. Pricing page claims “over 5 million” participants; the interview page says “over 3M”. The figures do not match	All plans, “limits apply”	Panel credits, Pricing, Interview studies
Interview studies	Moderated sessions with scheduling, built-in video or Zoom/Meet/Teams, transcripts and AI summaries	Enterprise	Interview studies
AI moderator	AI-run interviews driven by your goals and discussion guide. Image stimuli supported (up to five JPEG/PNG files per study)	“An add-on with our Enterprise plans”	AI moderator
AI study builder, MCP server	Drafts a study from stated goals; MCP server listed in the feature table	Enterprise	Pricing

Naming overlaps: “Maze” is a common word. Use maze.co (marketing), app.maze.co (product) and help.maze.co (docs). The help centre still uses a legacy maze.design sales email (Panel credits). Do not confuse it with unrelated products or with generic “maze” UX demos.

Pricing (re-check live): the public pricing page shows no prices. It tags features as “All plans” or “Enterprise”. A third-party tracker reports that the $99/month Starter plan was removed in June 2026, leaving Free and Enterprise (PricingSaaS). Maze’s help centre still documents a Starter upgrade flow and says pricing can differ by account, so check app.maze.co/upgrade. It also mentions a 30-day trial team for new accounts (Upgrade help). Free plan study and seat limits: not stated publicly on the official pages reviewed. Enterprise price: not stated publicly.

Panel costs (official): credits cost USD 5 each in bundles (50 for $250, 100 for $500, 500 for $2,500). They are valid for 12 months. Per response: unmoderated Core 3 credits, Extended B2C 5, Extended B2B 9; AI-moderated 5 (B2C) or 9 (B2B); moderated 25 (B2C) or 45 (B2B). Maze says these prices include incentives. Unmoderated panel studies are capped at 45 minutes estimated length. Maze recommends at least 10 participants for unmoderated studies (Panel credits). So 10 Core unmoderated responses cost 30 credits, roughly $150.

2. How it scores on the rubric

Criterion	Assessment	Evidence
Time-to-useful	Fast for unmoderated work. Results show up as participants complete the test; vendor quotes describe AI-moderated reports “within a few hours” (a testimonial, not a benchmark). Panel turnaround SLA: not stated publicly	Prototype testing, AI moderator
Fidelity (of evidence)	Good behavioural signal on tasks: paths, misclicks, first click. Weak on “why” unless you add open questions or pay for Enterprise interviews. The Usability Score is a vendor composite, so treat it as directional	Prototype testing
Non-engineer usability	High. No code, and the participant side is web-based. The real skill gap is research design: writing unbiased tasks and screeners. Maze’s “no researcher required” claim refers to the Enterprise AI moderator	AI moderator, Interview studies
Design-system / prototype fidelity	Strong for a solo builder’s stack: Figma prototype import, plus Figma Make, Lovable, Bolt and Replit prototypes; desktop, tablet and mobile formats. Fidelity depends on your prototype’s interactions	Prototype testing, Figma integration
Human gates	You set the goals, discussion guide, screeners and success elements. AI reports are editable before sharing. Maze claims “25 quality metrics” per AI conversation; that figure is unaudited. Nothing stops a solo builder from approving their own leading questions	AI moderator, Interview studies
Handoff	Shareable automated reports; reports can be embedded in Notion, FigJam or Miro; downloadable reports; MCP server on Enterprise. Direct ticket creation in issue trackers: not stated on the pages reviewed	Prototype testing, Pricing

Fit for this section (verdict: yes-with-caveats). A solo builder can define success criteria as Maze missions, buy 10 or more panel responses and get task-level evidence the same day, on the Free or entry tier. That is the honest core. The caveats: problem validation through conversation (AI moderator, interviews) is Enterprise; self-serve plan pricing is unclear and changing; and Maze measures your tasks, it does not challenge them. Pair it with a few of your own customer calls for the “is this the right problem” question.

3. FAQ

Can a solo builder validate a problem with Maze and no research team?

For solution and usability validation, yes: unmoderated prototype tests plus the panel. For problem discovery interviews, Maze’s answer is the AI moderator, which is an Enterprise add-on (AI moderator).

Does Maze import Figma prototypes?

Yes. It also supports Axure and AI-built prototypes from Figma Make, Lovable, Bolt and Replit (Prototype testing).

How much does Maze cost?

The official pricing page lists no prices. A third-party tracker reports Free and Enterprise only after June 2026 (PricingSaaS). Enterprise pricing: not stated publicly.

How much do Maze panel participants cost?

Credits are $5 each. An unmoderated Core response costs 3 credits; moderated B2B costs 45 (Panel credits).

Is the AI moderator on the free plan?

No. Maze says it is an add-on with Enterprise plans (AI moderator).

Can I test after shipping?

Live website and mobile app testing are listed, but marked Enterprise on the pricing page (Pricing). On lower tiers, test a prototype or a staging build imported as a supported prototype.

Is there an affiliate programme?

No public affiliate or partner page was found (maze.co/partners and maze.co/affiliates return 404). No affiliate programme was confirmed.

4. Sources

1. Maze (homepage)
2. Maze pricing and packages
3. Prototype testing
4. AI moderator
5. Interview studies
6. Surveys
7. Figma integration
8. How much do panel credits cost? (Maze Knowledge Base)
9. Hiring panel participants for AI Moderated testing (Maze Knowledge Base)
10. How do I upgrade my team? (Maze Knowledge Base)
11. Maze cuts Starter plan, leaving only Free and Enterprise tiers (PricingSaaS Pulse, third party)