UI Verify: visual testing for agents
UI Verify
UI Verify's remote MCP server: the agent reviews pull request UI changes, inspects diffs and accepts the intended baselines
- 27 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The agent accepts new baselines (accept_build, review_diff), and a wrong call can hide a real regression
- The service receives snapshots and a build of your interface, including page resources
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/ui-verifyDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent accepts new baselines (accept_build, review_diff), and a wrong call can hide a real regression
- The service receives snapshots and a build of your interface, including page resources
This is third-party code. Review the repository files before installing.
What it does
UI Verify checks your interface in CI: it renders components in the cloud, diffs them against the baseline from the default branch and reports a verdict as a GitHub check. Snapshots come from Storybook, from Playwright and Vitest tests through an SDK, from bun tests, or from finished screenshots of native and mobile projects. The remote MCP server at https://uiverify.ai/mcp gives the agent the review. list_projects and list_builds show projects and builds, get_build and get_pr_changeset describe what changed in a build or a pull request, and get_diff and render_diff_image return the actual pixels. review_diff sets a verdict on one story and accept_build on a whole build. The repository holds the client packages: the uiverify CLI, SDKs for Playwright, Vitest and Bun, and a set of skills. The uiverify check command renders only the stories you name while you edit and answers whether the change would pass CI, without moving any baseline.
Who it is for. Frontend teams with Storybook or Playwright tests who want to review visual changes in a pull request through Claude Code or Codex.
Good fit when
- A pull request touches dozens of components and the agent has to separate real regressions from cosmetic noise
- You want to accept new baselines in bulk after a deliberate redesign of buttons and forms
- While editing, the agent needs to know whether a story would pass in CI without waiting for a full build
- A mobile project produces Maestro or Detox screenshots that must be compared between builds
- A failed visual test is explained from the diff image, with no reliance on a changed-pixel count
Not a fit when
- You need logic or data checks: the service compares only the appearance of the interface
- You need a fully local run: rendering, diffing and judging happen on the service side
- The project has no CI and no Storybook, Playwright, Vitest or bun: you first need a way to produce snapshots
Example request
Show what my pull request changes in the UI, separate regressions from noise and suggest which baselines to acceptMCP
- Transport
- http
- Authentication
- OAuth
Limitations
UI Verify is a hosted service: the CLI needs a UIVERIFY_API_KEY from the dashboard at uiverify.ai. The MCP server is remote and signs in through the browser, once for every project and team you belong to.
The open repository holds only the client side: the packages build a bundle, upload it over HTTP and read back status. Rendering, diffing and judging run on the service. Plan terms are not described in the repository, so check the website. The agent's write rights follow your access in the project.
How to disable. Remove the server with claude mcp remove uiverify and, if needed, its line from .mcp.json. You can also revoke access in the UI Verify dashboard.
FAQ
How do I connect the UI Verify MCP server to Claude Code?
Run claude mcp add --scope project --transport http uiverify https://uiverify.ai/mcp, restart Claude Code, open /mcp, pick uiverify and sign in in the browser. The --scope project flag writes the server into the shared .mcp.json.
Which tools does the server have?
Ten: list_projects, list_builds, get_build, list_build_stories, get_pr_changeset, list_pr_stories, get_diff, render_diff_image, review_diff and accept_build.
What does uiverify check do?
While you edit, it renders only the stories you name (--target with a glob or exact id) and compares them with the real CI baseline. No check shows up on the pull request and no baseline moves.
How do I install the UI Verify skills?
With npx skills add uiverify/uiverify or as a Claude Code plugin: /plugin marketplace add uiverify/uiverify, then /plugin install uiverify@uiverify. On Gemini CLI an extension installs the same skills together with MCP.
Does UI Verify work with mobile projects?
Yes, by uploading finished screenshots: npx uiverify upload --screenshots ./screenshots. Each image is keyed by its path inside the folder, so a screen maps to the same baseline from build to build.
Discussion
No questions or reviews yet.
Related
MCP serversA Claude Code plugin that verifies a change against its plan: boots your stack, runs acceptance criteria, reports evidence.
Checks on-screen text with local OCR: an MCP server and SDK for Playwright and Cypress that catch wrong, faint or missing text
Snapshot tests for AI agents: records tool calls today and tells you when they silently change after an edit
MCP server, agent and skills for end-to-end UI tests where the agent drives the computer through vision on cloud machines
An MCP server that checks whether the agent is really done: drives the real app and returns pass, fail or couldn't tell with the file and line
A skill that tests your agent's code in a real browser: reads your changes, builds a test plan and runs it through Playwright with your own cookies
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| UI Verify: visual testing for agents (this card) | yes | OAuth | 27 | |
| Playwright MCP | yes | not required | 39,570 | drives a browser and a page, while UI Verify compares interface snapshots with a baseline |
| Playwright Skill | no | 3,287 | teaches the agent to write Playwright tests, while UI Verify reviews the outcome of their visual comparison |
README in short
The README explains four capture paths, a quickstart with npx uiverify upload, a package table, skill installation and MCP setup. Examples sit alongside for each path: Storybook, Vitest, Playwright, Bun and React Native. The license is MIT.