Run it yourself
From clone to first finding.
The fastest path is a CLI scan of a local directory — only an LLM key is needed, no Supabase and no GitHub App. You bring your own key, you keep the findings, and only the files being scanned are sent to your provider.
1. Clone & install
git clone https://github.com/Nikolaospet/themida.git cd themida pnpm install
2. Configure your LLM provider
Copy the example env file and set your provider plus key. Themida works with Anthropic, OpenAI, or any OpenAI-compatible endpoint.
cp .env.example .env.local # in .env.local — pick one: LLM_PROVIDER=anthropic ANTHROPIC_API_KEY=sk-ant-… # …or an OpenAI-compatible endpoint: LLM_PROVIDER=openai OPENAI_API_KEY=sk-… OPENAI_BASE_URL=https://openrouter.ai/api/v1
3. Scan
# scan this clone, or any local directory
pnpm dev:scan --path . Useful options
# restrict to specific frameworks pnpm dev:scan --path ./my-repo --frameworks gdpr,owasp # write a SARIF report for GitHub Code Scanning pnpm dev:scan --path ./my-repo --format sarif --out themida.sarif
Want the full web UI?
The dashboard adds repo connection, scan history, framework selection, and PDF / SARIF export from the browser. It needs a local Supabase instance and (optionally) a GitHub App. Follow the local setup guide to run the whole stack.
Themida is alpha software — treat findings as input to human review, not a certified audit.