Run it

Ten minutes, your own key, the real thing.

This runs the real vertical slice — risk assessment, test generation, human review, and retrieval memory — against whatever LLM you point it at. Everything past that (execution, defect triage, release readiness) is roadmap, not yet runnable. See /roadmap.

# Backend
cd services/api
uv sync
export LITELLM_BASE_URL=https://api.openai.com/v1   # or any OpenAI-compatible endpoint
export LITELLM_API_KEY=sk-...
uv run uvicorn main:app --reload

# (optional) persistent demo data across restarts
export KG_SQLITE_PATH=./qa_os.sqlite3

# Frontend
cd apps/web
npm install
npm run dev
# open localhost:3000 — type a requirement, run the pipeline, approve a
# test, then click "View Graph" to see it live in the graph explorer

Found something worth fixing, or want to wire up Module 5 or 6? Issues and pull requests are genuinely welcome — see the repo for exactly where things stand.

View on GitHub See what's implemented

Elsewhere

More projects, engineering philosophy, and the rest of the story live at buildwithdaksh.com.