One step in your pipeline scans the site you just deployed and fails the build on the problems you choose, so a broken sign-up never reaches production.

GitHub Actions gets a summary, annotations and step outputs with no setup. GitLab, CircleCI, Jenkins, Bitbucket, Buildkite and Azure Pipelines read the JUnit report.

Setting up CI in ValiqAI, then the GitHub Actions run a scan fails, and back to the results.

How it works

You deploy. To staging or a preview URL.
The pipeline scans it. npx -y @valiqai/cli scan runs against the address you just deployed and waits for the result.
The build passes or fails. On blockers by default (--fail-on to change it), with the results linked from the run.

Set up

In ValiqAI, open a team’s Integrations › CI.

Make a team API key and store it as VALIQAI_API_KEY in your CI secrets.
Add the step for your CI system (it is on the setup page).
Point it at the address you deploy.

Plan

CI is on the Pro plan (Pro, Teams and Enterprise). A connection made before a plan change keeps working.

Step-by-step guides for each CI system are in CI guides.

With AI tools

With the ValiqAI MCP server added to your AI coding tool, you can ask:

“Add a ValiqAI scan step to our GitHub Actions deploy workflow.”
“Why did the ValiqAI step fail on the last deploy? Show me the blocker.”