Notable releases, newest first, with the day they went live. Smaller fixes land every day and are not listed.
The CLI, the MCP server and both monitoring SDKs are published on npm under MIT, with their source on GitHub.
Opt-in replay in checkd-browser 1.1.0. It records around errors by default, masks every input in the browser and plays back on the issue page with error, click and request markers.
Any issue can be explained in plain English: what is broken, the likely cause, who is affected, where to look, a suggested fix and how sure it is. From there, open a fix pull request on GitHub.
Browser and Node.js SDKs send the errors, failed requests and slow pages real users hit. They are grouped by root cause into issues, next to the findings from test runs, scans and load tests.
Record a journey in a browser shown inside Check’d, or describe it in plain English. Journeys replay with several locators per step, self-heal with your approval, and fresh test data on every run.
Paste a link to published API docs and Check’d finds the definition behind it, on Swagger UI, Redoc, Scalar, Stoplight, Postman, Apidog, SwaggerHub, ReadMe, Mintlify, GitBook, Bump.sh, Zuplo, Docusaurus or any docs site.
Sign in from the terminal by approving a code in the browser, import specs, start runs with live progress and CI exit codes, and connect Claude Code, Cursor or VS Code through MCP.
A new name and the Ch’D mark across the app, emails, reports and the share image, on getchecked.twelveai.app.
Starter $20, Business $100 and Team $200 a month, with usage and extra credit in US dollars. The app moved to a calmer workspace layout with tables and metric strips, and the marketing site was rebuilt.
When a test hits something only you know, such as a staff sign-in or a one-time code, the run asks you and carries on. Projects can be archived and restored.
Choose Critical, Standard or Deep for each run, and watch suggested fixes appear on the board as a run finishes.
Requirements traceability, release gates, a written test plan, release certificates, OWASP API Top 10 mapping, contract drift, Gherkin export, compliance packs and a test board.
Runs on pull requests with a summary on the PR, AI fix pull requests, GraphQL support, findings sent to Linear or Jira, shareable report links and the free site scan.
Website scans with a site health score, and load tests that stop by themselves before a firewall flags your address.
Large runs split across parallel test agents, and load tests scale up to 30,000 virtual users on the Team plan.
Start free, no card needed.