AI-generated E2E tests from every commit — for web and Android.

AutoSpec AI is a set of GitHub Actions that watch your code changes, understand what user-facing behavior changed, and generate production-quality E2E tests that match your existing test style. No infrastructure to manage. No test environment to configure. Just add the action and start shipping tests.

Commit / PR  →  Diff Analysis  →  LLM Test Planning  →  Code Gen  →  PR with Tests

Each action analyzes your diff, plans which user flows need testing, generates test code with severity tags (@sev1 through @sev4), and opens a PR — all within your existing GitHub Actions workflow.

Generate Playwright E2E tests for web apps. Built-in trace capture, API mock generation, visual regression, and accessibility scanning.

- uses: autospec-ai/playwright@v1
  with:
    llm_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
    base_url: 'http://localhost:3000'
  env:
    GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

Generate Maestro (YAML) or Espresso (Kotlin) tests for Android apps. Auto-detects project context from build.gradle and AndroidManifest.xml. Classifies changed files by UI impact and generates tests for affected screens.

- uses: autospec-ai/android@v1
  with:
    llm_api_key: ${{ secrets.ANTHROPIC_API_KEY }}

Maestro is the default — declarative YAML tests that work against any APK without injecting dependencies into your build. Espresso is available as an opt-in for teams with existing infrastructure.

• Two-phase LLM generation — plans tests first, then generates code
• API Mock Generation — detects HTTP patterns and generates framework-appropriate mocks
• Severity tagging — sev1 (critical flows) through sev4 (edge cases)
• Accessibility — WCAG scanning (web) and content description validation (Android)
• Screenshot comparison — visual regression baselines for both platforms

Supports Anthropic, OpenAI, and any OpenAI-compatible provider (Ollama, Together, Groq).