🔨
FORGEQA
ARCHITECTURE
Self-hosted AI → Playwright Python tests
Phase 1 — test generation + human review (execution & RAG on the roadmap)
LOCAL LLM LAYER
LiteLLM → one code path
- Dev: LM Studio @ localhost:1234
- Deploy: Ollama in Docker
- Cloud fallback off by default
PAIRS WITH TESTMCP
Plan → Code
- TestMCP predicts what to test from a PR
- ForgeQA generates how to test it in Playwright
Your infra. Your models. Your test code.
Built by Ramona Bonitatis · Phase 1 MVP · github.com/ramonacraft/forgeqa