--- title: "Let your coding agent verify its work." description: "Connect Cursor, Claude Code or Codex to Momentic's MCP server. Agents write YAML tests into your repo, run them in a real browser, and investigate failures." canonical: "https://momentic.ai/coding-agents" last-updated: "2026-11-12T09:10:03Z" --- # Let your coding agent verify its work. URL: https://momentic.ai/coding-agents platform / coding-agents Give Cursor, Claude Code, or Codex access to the running app through MCP. Your agent can test a change, inspect failures, and save the test for future runs. Copy agent prompt Export a web app API key as MOMENTIC_API_KEY or use CLI login. [Try for free](https://app.momentic.ai/signup) [Get a demo](/sales) Cursor · Claude Code · Codex · OpenCode · GitHub Copilot ## Close the loop between code and behavior. YAML Tests in your repository Live testing Your agent can preview steps on the running app through MCP. Failure details Inspect failed-step results and available screenshots. Saved tests Keep reusable YAML tests beside your application code. Web + mobile Use separate MCP tools for web, iOS, and Android. ## From a code change to a verified flow. Give your agent the skills and tools to test the app. Review the expected outcome with the code change. 01 ### Register the server Connect your coding agent to Momentic's MCP server. Give it tools to inspect the app, preview steps, and run tests. 02 ### Install the skills The web and mobile CLIs install the same three skills. Your agent uses product references to write tests, specify a change, or investigate failures. 03 ### The agent writes and runs the test The agent tests a live browser or device, saves the YAML, and inspects the result. Review the diff, then run the same test in CI. ## MCP tools for end-to-end verification. Let the agent run the app, inspect the results, and correct its work. Keep the resulting tests in your repo for CI. ### Write tests against the running app Your agent opens the app through MCP, tries each step, and saves it after checking the result. The agent sees the real page before it saves a step. ### Define the expected behavior before coding Use testing skills to define expected behavior before changing the code. Your agent writes the test, then implements the change. The agent writes the check first, then makes it pass. ### Give the agent evidence to investigate Your agent reads past runs and step evidence to find where behavior changed. It uses that history to investigate the failure. The agent reads the run history before it edits a test. ### Web, iOS and Android Your agent can author and run tests against browsers, iOS simulators, and Android emulators. All three use the same simplified YAML format. Same format on all three platforms. ### Repair tests from run evidence The agent inspects the failed step and proposes a test change. Rerun the test and review the diff before keeping the repair. Review the repair after the test passes. ### Review the test with the code change Every test an agent writes is YAML in your repo, so it goes through code review and the CLI runs it in CI. Agent work lands in a pull request. Engineering teams that let agents keep the tests current. Engineers stay in the editor, and the tests still land in the repo with the code change. “Momentic gave us a fast and reliable way to validate Poe.com's AI responses, even when they weren't deterministic.” Momoko F. Head of Product Operations, Quora 30 min daily test execution, down from 7 hours 500+ manual test cases replaced 100% critical tests created in one month [30 min daily test execution, down from 7 hours "Momentic gave us a fast and reliable way to validate Poe.com's AI responses, even when they weren't deterministic." Momoko F. Head of Product Operations, Quora](/customers/quora) [8x increase in release cadence "It's like giving someone your QA checklist and watching them execute it for you!" Sriram S. Engineering Lead, Source Control, Retool](/customers/retool) [80% faster release cycles "With Momentic, we've caught bugs that would have eluded even our most diligent internal tests." Alex C. CTO, GPTZero](/customers/gptzero) [6x faster end-to-end test creation "We've already seen a 30% decrease in production incidents thanks to Momentic's automated testing." Hanna K. Head of QA, CoverGo](/customers/covergo) [85% reduction in production incidents "Momentic gives us reliable end-to-end coverage, so we can focus on features instead of maintaining tests." Alec H. Staff AI Engineer, Mutiny](/customers/mutiny) [Browse all](/customers) ## Test web, iOS, and Android. Use the same simplified YAML format across web, iOS, and Android. [01 Web End-to-end browser testing on Chromium and Chrome.](/web-app-testing)[02 iOS Native iOS apps and SwiftUI, on hosted or local simulators.](/ios-testing)[03 Android Run Android APKs on local or hosted emulators.](/android-testing) ## Coding agents, answered. What is the Momentic MCP server? MCP gives coding agents Momentic tools. Local momentic and momentic-mobile servers author and run repository tests. momentic-hosted at https://mcp.momentic.ai provides docs and Mo tools over HTTP with OAuth. Hosted MCP does not replace the local test server. Which agents work with it? Cursor, Claude Code, Codex, Devin, OpenCode and GitHub Copilot. Run npx momentic install-mcp --client claude-code for web or npx momentic-mobile install-mcp --client claude-code for mobile. Replace claude-code with your agent's MCP client name. Each registers its local server and momentic-hosted. How do I set it up in Claude Code? Run npx momentic skills --target claude-code and npx momentic install-mcp --client claude-code. For mobile, use momentic-mobile in both commands. If you use both CLIs, upgrade them together. Restart Claude Code, approve the local server in /mcp, and sign in to momentic-hosted with OAuth. Use either the Momentic plugin or install-mcp, not both. Plugin users should decline the CLI skills prompt or use init --config-only. What are Momentic skills? Markdown instructions for coding agents. The shared web/mobile bundle includes momentic-test for authoring, momentic-spec for tests before product changes, and momentic-maintain for failure investigation. Install with npx momentic skills --target claude-code or npx momentic-mobile skills --target claude-code, replacing claude-code with your agent's target. The plugin adds mo-qa; npx -y qa skills --target claude-code installs only mo-qa. Does the agent need an API key? Local MCP needs an API key. Launch your editor from a shell that exports MOMENTIC_API_KEY, or use the CLI login command. Hosted MCP uses OAuth, not an API key. Where does the agent save the test? As a YAML file in your repo, next to your code. You review it in a pull request. Can the agent run the test it just wrote? Yes, through the same server, in a real browser. The CLI runs that same file in CI later. Our coding agent already runs our Playwright suite. What changes? Your agent stops maintaining the runner. Momentic can find a moved control again and retry when a temporary obstruction blocks a step. MCP returns the failing step with its evidence. The [build vs buy calculator](/comparison/momentic-vs-build-your-own) prices the suite your agent keeps. Is that cheaper than having the agent maintain a Playwright suite? Not for every suite. Compare Momentic usage and CI costs with the agent time your current suite spends on maintenance. Runtime recovery and classified evidence can reduce work, but costs depend on your measured usage. What does the agent get back when a test fails? The failing step and its evidence: before and after screenshots, a DOM trace, network requests, console output, execution order, and video when recording is on. momentic-maintain also reads branch run history to find where behavior changed. Does the agent remember how our product works? Momentic supplies product context. Knowledge base entries guide AI steps, and Memory supplies prior locator and assertion decisions during authenticated runs. Memory depends on caching and expires after 30 days of inactivity. Does this replace the visual editor? No. The editor writes the same YAML, so one suite can hold tests from engineers, agents and the editor. Which CLI version do I need? Use the latest momentic CLI for web or momentic-mobile for mobile. Both include the shared skills installer and install-mcp. What is the difference between Playwright MCP and Momentic's MCP server? Playwright MCP drives a browser and writes Playwright code, so authoring-time selectors and strings become the test. Momentic previews each step on the live page, saves working steps as YAML, and replays them with step caching and auto-heal. Can an agent run tests and read the results over MCP? Yes. The agent runs a test through the same server, and the momentic-maintain skill reads the run history, the saved classifications and the step evidence for a branch. Does MCP keep our tests in version control? Yes. The agent writes a YAML file into your repo, so a test change arrives in a pull request and CI runs it on the next commit. Can a coding agent run tests in CI over MCP? No, and it does not need to. MCP is a local server for your editor. CI runs npx momentic run, and the exit code gates the merge. Still have additional questions? ## Try Mo on your app. Mo is an AI QA engineer. It tests your running app and files each bug with repro steps. Free to start, no credit card. [Try for free](https://app.momentic.ai/signup) [Get a demo](/sales)