Changelog¶
v1.0.0 — 2026-05-30¶
Initial production release.
Features¶
10-agent AI pipeline
RequirementAgent— parses user stories into structured modules and risk areasTestCaseAgent— generates UI, API, and edge case test cases with RAG enrichmentVerificationAgent— coverage analysis, duplicate detection, gap identificationSeleniumAgent— generates executable Python Selenium scripts per test caseAPIAgent— generates httpx HTTP test suites with full assertionsExecutionAgent— MOCK / LOCAL / GRID execution with HITL approval gateBugAgent— root cause analysis with RAG correlation and clusteringHealingAgent— 7-level self-healing locator recoveryReportAgent— GO / GO_WITH_RISK / NO_GO release decision engine
CLI (testpilot)
testpilot init— interactive setup wizard, writestestpilot.yamltestpilot run— full pipeline with exit codes 0/1/2testpilot analyze— test case generation onlytestpilot bugs— AI + RAG bug analysistestpilot report— release decision from results JSONtestpilot dashboard— Streamlit visual dashboard
Python API
run_pipeline(),analyze(),analyze_bug(),generate_report(),configure()- All Pydantic v2 schemas re-exported as public API surface
Storage
- SQLite via SQLAlchemy (6 tables, PostgreSQL-ready via URL swap)
- ChromaDB embedded RAG (4 collections, all-MiniLM-L6-v2)
Observability
- LangSmith tracing (optional, free tier)
- Loguru structured logging
Deployment
- Streamlit Cloud (live at ai-testpilot-x.streamlit.app)
- Docker Compose for Selenium Grid
pyproject.tomlwith optional extras:ui,selenium,playwright,docs,all
CI/CD
- GitHub Actions workflow examples
- GitLab CI example
- Jenkins Declarative Pipeline example
Known limitations (V2 roadmap)¶
RiskAgent— deferred to V2 (stub present)OptimizationAgent— deferred to V2 (stub present)- PostgreSQL — supported via
DB_URLenv var, not tested in CI yet - Playwright execution — stub only, not wired to execution graph
- PyPI publish — package installable from source, PyPI release pending