# How to Run AI Tests on Every Pull Request & in CI/CD > Automate testing on every pull request and in your CI/CD pipeline with AI agents – test the change in the PR thread, gate deploys, and run against preview URLs. Source: https://qa.tech/use-cases/pr-cicd-testing --- A GitHub or [GitLab](https://docs.qa.tech/configuration/gitlab) bot installs into your repo; when a PR opens (or you tag it), it reads the PR context, detects the [preview](https://docs.qa.tech/core-concepts/applications-and-environments#preview-environments) deployment your pipeline built and runs exploratory plus relevant regression tests against it, then posts results back to the PR thread – so issues are caught before merge, without leaving the developer's workflow. [Book a demo](https://qa.tech/demo) Sub-use-cases **Covers** PR bot testing, preview-URL testing, CI/CD pipeline gate (blocking or reporting), post-deploy smoke, scheduled pipeline regression, GitHub App (automatic trigger on PR open or label), GitHub Actions Change Review Action (multi-app PRs, label-based triggers, custom pipelines), CLI (\`qatech run\`, \`qatech tunnel\`) for local development and non-standard pipelines and API / webhook trigger from any other CI system. 1. 01 ## What gets tested on a pull request PR-level testing of the change, preview-environment (Vercel/Netlify/Railway) runs, pipeline quality gates on deploy, and post-deploy production smoke. 2. 02 ## How does AI test a PR before merge? [GitHub Actions](https://docs.qa.tech/configuration/github-actions) and [GitLab](https://docs.qa.tech/configuration/gitlab) CI are native; other systems trigger via API or [webhook](https://docs.qa.tech/configuration/ci-cd-integration). The bot auto-detects the PR's preview URL, runs the plan, and returns pass/fail with screenshots and video to the thread or pipeline. 3. 03 ## When tests run in the pipeline On every PR and at each pipeline stage – a smoke plan on merge, full regression on release candidates. 4. 04 ## Who runs PR and CI/CD testing Dev and platform teams who want quality gates earlier, without leaving their tools. 5. 05 ## How QA.tech helps Code review, whether a human or a tool like CodeRabbit, catches logic errors, not user-facing regressions; the [CodeRabbit alternatives](https://qa.tech/compare/coderabbit-alternatives) we compared all share that limit. QA.tech moves testing to the moment of the PR, so a broken checkout or login is caught before it reaches main. [API test plans](https://qa.tech/product/api-testing) run in the same gate, so a broken endpoint contract is caught alongside the broken checkout. 6. 06 ## Ways to integrate QA.tech plugs into your pipeline through several entry points: - \*\*[GitHub App](https://docs.qa.tech/configuration/github-app)\*\* – automatic trigger on PR open or label. - \*\*[GitHub Actions](https://docs.qa.tech/configuration/github-actions) Change Review Action\*\* – a standalone action for when the App's auto-trigger isn't flexible enough: multi-app PRs, label-based triggers, custom pipelines. - \*\*CLI\*\* (\`qatech run\`, \`qatech tunnel\`) – for local development and non-standard pipelines. - \*\*API / [webhook](https://docs.qa.tech/configuration/ci-cd-integration)\*\* – trigger from any other CI system. ▒ ░▒ ▒ ▓░░ ░▒▓▓░▓░░ ░ ░░ ░░░▒░▒▒░░ ▒░▒░ ░ ░░░░ ░░▓░░▒░ ░░▒█▒ ░ ░▒█ ▓▒ ░░ ▓▒░▒█ ▒ ░░▓ ░░▓░▒ ░░░░░ ░ ░ ░▒▒ ░▒░▓░ ░▒ ▓ ▒▒░░ ▓ ░░░░░░ ▓▓░ ░ ░ ░▓ ░ ▓█ ░▒▒░ ░▒ ▒▒▓ ░ ░░ ░░░░▓ ▒ ▒ ░░▒ ▒▒░ ░▒ ▒▒ ▓░░▓ ░▒▓▓░▒░░█░░▒▒░ ▒░ ▒▓▒ ░░▓ ▓ ▒▒ ░▒▒░░ ░ ▒ ▒░░░░ ## Companies running pull request & ci/cd testing with QA.tech - [Tests every pull request in GitHub and saves 390h of testing per quarter.](https://qa.tech/case-studies/how-pricer-transform-qa-with-qa-tech) - [![4C Strategies](https://qa.tech/assets/4c-strategies-CVLIdhMC.svg)Tests every merge request on a preview deploy through self-hosted GitLab.](https://qa.tech/case-studies/4c-strategies-replaces-manual-testing-with-ai) - [Wired QA.tech into a custom Azure DevOps pipeline via API.](https://qa.tech/case-studies/bizbrains-ui-testing-automation) - [Runs shift-left PR tests in GitHub.](https://qa.tech/case-studies/smartlinx-ai-qa-replacement) - [![Crystal Intelligence](https://qa.tech/assets/crystal-intelligence-B-g-38p_.svg)Uses QA.tech as a PR quality gate.](https://qa.tech/case-studies/crystal-intelligence-speeds-up-regression-testing) - [![Airpelago](https://qa.tech/assets/airpelago-1y9XaVff.svg)Triggers the full E2E suite from a GitHub commit message.](https://qa.tech/case-studies/airpelago---scaling-drone-operations-with-reliable-end-to-end-testing) FAQ ## Common questions Does QA.tech test the PR's preview deployment automatically? Yes – the bot detects the preview URL and runs against it, no per-PR config. Can QA.tech block a deploy on failure? Yes – wire it as a [blocking check](https://docs.qa.tech/configuration/ci-cd-integration); non-critical tests can report without blocking. Which CI systems are supported? GitHub and GitLab natively; Bitbucket, Azure DevOps and others via API/webhook. Related use cases - [ API Regression Testing You group API contract checks into a plan written as plain-language goals – "GET /orders returns 200 with a list, and every item has an id and a status" – and an AI agent runs them against your API environment on every deploy. The agent makes the requests, writes and executes its own validation code, and returns a verdict with the full request and command trace. Read ▒ ▓ ▓ █░░░ █ ▒ ▒░▓░ ▒▓░ ░░▒░▓░▒▒▓░▓░░ ▒░ ▒ ▓ ░ ░▒ ░░░ ░▒▒▓░░ █░░░░ ▓░░▒░░ ▒▒ █░ ░ ░▒ ░░ ▒ █░ ▒░░░█░ ░ ▓▒▓▓ ▓](https://qa.tech/use-cases/api-regression-testing) - [ Automated Regression Testing Regression testing re-checks that existing features still work after a change. To automate it with AI, you group tests into a regression plan written as plain-language goals, and agents run the whole suite in parallel on every deploy. Many teams wire this into PR testing so the suite runs on every pull request preview. A 50-test suite that took hours by hand finishes in around ten minutes, and the tests don't need rewriting when the UI shifts. This is the workflow behind our automated web testing product. Read ░ ▒░░▒░░░░ ▓▒▓░ ▓░░ ░░▒░░█ ░▒░ ░▒░▓░ ░ ░ ░▒ █ ░ ░ ░█▒▓ ░ ▓ ░█ ░ ░░ ▓░▒ ░▒▓ ░░░░░▓░ ░░░ ▓░ █ ░░░ ▒░░▓▒▒░ ▒░▒░ ▒▒▒](https://qa.tech/use-cases/automated-regression-testing) - [ End-to-End Testing End-to-end testing verifies a complete user journey works from start to finish, the way a real user experiences it. With AI you describe the journey as a goal and an agent carries it out – logging in, navigating, filling forms, verifying the outcome – across your whole app. QA.tech runs these in parallel and re-navigates when the UI changes, so long flows don't collapse on a renamed button. In practice most teams run these journeys as part of automated web testing on every release. Read ░▓▓▓░ ▒▒░░ ░▒ ░ ░░ ░░░░░░▒░░▒░ ░░█ ░░▒ █ ░█ █ ▒ ░░ ▓▒▒ ░░░░ ░░▓░▓▒░▓░▓ ░ ▓ █░▒▓ ▒ ░▓░ ░░░░▒ ░░ ▒▒▒ ▒░ ▓░░ ░░░](https://qa.tech/use-cases/end-to-end-testing) [← Previous Notifications & Activity Feed Testing ](https://qa.tech/use-cases/notifications-testing)[Next → QA for AI-Generated Code & Agentic SDLC (MCP) ](https://qa.tech/use-cases/ai-generated-code-testing) ## Your team moves fast. Can your testing keep up? QA.tech agents test your product autonomously, so moving fast never means shipping broken. See how it works in a 30-minute demo. [Get a demo](https://qa.tech/demo)