UserTesting MCP honesty matrix — create and launch studies vs FDT create_github_issue / file:line
UserTesting's MCP marketing page (usertesting.com/platform/mcp-server) says: "Use UserTesting's MCP server to recruit participants, create studies, and launch tests from Claude, ChatGPT, Figma Make, and other AI clients." The same page FAQ answers "What can teams do with UserTesting's MCP server today?" with "recruit participants, create studies, and launch tests from AI clients." Access is "Limited early access." Recruiting "is powered through UserTesting and User Interviews."
The Help Center open-beta article (help.usertesting.com/hc/en-us/articles/38351574748573-Introducing-the-UserTesting-MCP-Open-beta) is more precise than the marketing verb "create studies." It says the connector lets an AI client "describe a test in plain language, create it as a draft, launch it to participants, find your existing tests, and read the AI summary for a completed test." All clients use https://ai.usertesting.com/mcp. The published capability table covers listing accounts/workspaces, listing saved panel audiences, turning a description into tasks and questions, creating the test as a draft (link back to the test builder), launching an existing draft (permission + workspace approval), checking approval, searching tests, and reading the AI-generated summary. "Creating and launching a test are always two separate steps. Describing a test never launches it." Companion: help.usertesting.com/hc/en-us/articles/38352231885725-Create-and-launch-a-test-with-the-UserTesting-MCP-Open-beta.
That is a real write surface inside UserTesting — and it is not a ship-to-code surface. Nothing on the marketing page or the open-beta Help Center publishes create_github_issue, create_pr, open_pr, a codebase/git write, or React component file:line. Feedback DevTools foils with first-party create_github_issue / create_tracker_issue / mark_feedback_resolved plus capture-time source maps (component + file:line) handed to your local agent. We do not claim server-side auto-opened or auto-merged PRs. Zero UserTesting product integration ships in FDT today. Distinct from Lyssna#81, whose FAQ is read-only and cannot create studies — do not reopen Lyssna#81 / Pendo#80 / Contentsquare#79 / Dovetail#78 / FullStory#77.
Feedback DevTools is our product. UserTesting MCP marketing from https://www.usertesting.com/platform/mcp-server (recruit / create studies / launch tests; limited early access). Open-beta capabilities from https://help.usertesting.com/hc/en-us/articles/38351574748573-Introducing-the-UserTesting-MCP-Open-beta and create/launch steps from https://help.usertesting.com/hc/en-us/articles/38352231885725-Create-and-launch-a-test-with-the-UserTesting-MCP-Open-beta — lastChecked September 30, 2026 ~8:30am Asia/Jerusalem. Please verify before purchase.
Feedback DevTools vs UserTesting MCP
| Feedback DevTools | UserTesting MCP | |
|---|---|---|
| Positioning | Capture + create_github_issue / create_tracker_issue / mark_feedback_resolved + local-agent file:line | Research MCP — recruit participants, create studies/draft tests, launch tests, read AI summaries from Claude / ChatGPT / Figma Make |
| Published MCP write surface | create_github_issue · create_tracker_issue · mark_feedback_resolved + local-agent file:line | Open beta: create a test as a draft, launch that draft to participants (separate step), list saved panel audiences — research writes inside UserTesting, not GitHub |
| Marketing claim (recruit / create / launch) | N/A — FDT does not market study creation or participant recruiting | usertesting.com/platform/mcp-server: "recruit participants, create studies, and launch tests" from Claude, ChatGPT, Figma Make; limited early access |
| create_github_issue / create_tracker_issue / create_pr | Yes create_github_issue / create_tracker_issue + capture-time file:line; no server-side create_pr | No create_github_issue / create_tracker_issue / create_pr / open_pr / codebase / git write on the UserTesting MCP surface |
| Server-side auto-open / auto-merge PRs | No — we do not claim server-side auto PRs | No — published tools create and launch UserTesting tests; they do not open or merge GitHub PRs |
| Source maps / file:line | Yes — capture-time React component + file:line | Test builder link, tasks/questions, AI summary with links back to source tasks — research evidence, not React component file:line |
| UserTesting in FDT today | Zero — no UserTesting product integration ships in Feedback DevTools today | First-party remote MCP (https://ai.usertesting.com/mcp) in open beta / limited early access |
| Plan gate | Free plan forever + Pro $9/mo per project — MCP ticket tools without a research-suite Enterprise SKU | Open beta: account must be enabled (Generative AI features); marketing page is limited early access — not a $9 MCP SKU |
UserTesting MCP marketing verified on usertesting.com/platform/mcp-server; open-beta create-draft / launch / AI-summary capabilities verified on help.usertesting.com/hc/en-us/articles/38351574748573-Introducing-the-UserTesting-MCP-Open-beta — lastChecked September 30, 2026 ~8:30am Asia/Jerusalem
Where UserTesting MCP is genuinely strong
Credit where due — UserTesting MCP is a real research unlock if:
- You already run UserTesting and want Claude, ChatGPT, or Figma Make to turn a plain-language description into a draft test, then launch that draft to participants without leaving the AI client (open beta; create and launch are separate steps).
- Recruiting from saved panel audiences, with User Interviews in the same study-creation flow, matters more than React component file:line.
- Reading the AI-generated summary of a launched test, with links back to the source tasks, is the result surface you want inside the assistant.
Where Feedback DevTools is more honest about create_* / ship
- Create/launch studies is real — and it is not ship-to-code. Marketing (usertesting.com/platform/mcp-server) says recruit, create studies, and launch tests. Help Center open beta (article 38351574748573) says create a draft and launch it as a second step (lastChecked September 30, 2026 ~8:30am Asia/Jerusalem). Neither page publishes create_github_issue, create_pr, or file:line.
- Tickets with file:line. create_github_issue and create_tracker_issue open GitHub/Linear tickets whose bodies carry component + file:line from capture. UserTesting's published MCP surface writes UserTesting drafts and launches — a GitHub issue or PR, if any, is a separate coding agent, not a UserTesting MCP tool.
- Local-agent path. The same source map is handed to your coding agent; a PR, if any, is that agent's git workflow. FDT does not claim server-side auto-open or auto-merge.
- Capture-time source maps. Every item pins the React component and file path:line so the agent starts from a known path instead of a study ID or an AI summary alone.
- Not the Lyssna foil. Lyssna#81 is a read-only studies FAQ (cannot create or edit studies). UserTesting MCP can create drafts and launch tests. The honesty gap here is GitHub/code, not "can they write inside their own product." Do not reopen Lyssna#81 / Pendo#80 / Contentsquare#79 / Dovetail#78 / FullStory#77.
- Zero UserTesting in FDT today. This foil is honesty about surfaces — Feedback DevTools does not ship a UserTesting connector; it ships create_github_issue / create_tracker_issue + local-agent file:line.
Pricing shape (not the foil) (as of September 30, 2026)
- UserTesting: enterprise research platform; MCP is open beta / limited early access, not a published $9 SKU (sources: usertesting.com/platform/mcp-server; help.usertesting.com/hc/en-us/articles/38351574748573-Introducing-the-UserTesting-MCP-Open-beta)
- Feedback DevTools: Free (unlimited collection) · Pro $9/mo per project (source: feedback-devtools.com/pricing)
MCP honesty is about capability claims, not sticker shock — FDT Pro is $9/mo per project vs UserTesting as a research suite whose MCP creates and launches studies but does not publish create_github_issue / create_pr / file:line.
Who should pick which?
- Choose UserTesting MCP if your job is recruiting participants and creating/launching UserTesting studies from Claude, ChatGPT, or Figma Make, then reading the AI summary back in that client.
- Choose Feedback DevTools if you want visual feedback with file:line on capture, first-party MCP tools that create GitHub/Linear tickets (create_github_issue / create_tracker_issue), and a local-agent fix path — without reading "create studies and launch tests" as a GitHub write or a source map.
Frequently asked questions
What does UserTesting MCP create and launch?
Marketing (https://www.usertesting.com/platform/mcp-server, lastChecked September 30, 2026 ~8:30am Asia/Jerusalem): recruit participants, create studies, and launch tests from Claude, ChatGPT, and Figma Make (limited early access). Open beta Help Center (https://help.usertesting.com/hc/en-us/articles/38351574748573-Introducing-the-UserTesting-MCP-Open-beta): create the test as a draft, then launch that draft in a separate step; also list workspaces and saved audiences, search tests, and read the AI summary. Server URL https://ai.usertesting.com/mcp. Describing a test never launches it.
Can UserTesting MCP create GitHub issues or pull requests?
No create_github_issue / create_tracker_issue / create_pr / open_pr / codebase / git write / file:line on the UserTesting MCP surface (usertesting.com/platform/mcp-server + help.usertesting.com open beta; lastChecked September 30, 2026 ~8:30am Asia/Jerusalem). Writes are UserTesting drafts, launches, and AI summaries. Feedback DevTools foils with create_github_issue / create_tracker_issue + local-agent source map and does not claim server-side auto-open or auto-merge. Distinct from Lyssna#81 / Pendo#80 / Contentsquare#79 / Dovetail#78 / FullStory#77.
How is this different from Lyssna MCP?
Lyssna#81 FAQ is read-only — it cannot create or edit studies. UserTesting MCP open beta can create a draft study/test and launch it. Neither publishes create_github_issue, create_pr, or file:line. Do not reopen Lyssna#81.
Try Feedback DevTools free — create_github_issue / create_tracker_issue with file:line, not a research-suite study launcher.