09/07/2026
Turn any incident — a note, an alert, a log — into a classified, audit-ready report. All on your own machine, so sensitive data never leaves the building.
✅ Grounded risk scoring — every severity comes with a transparent rationale (rule layer + AI, reconciled). Defensible, not a black box.
✅ Human-in-the-loop — the AI drafts; you review and edit every field before export.
✅ Local-first, cloud-optional — runs fully offline by default, or point it at any hosted model (OpenAI, Azure, Gemini, Groq, Together) with one config change. Zero code edits.
Built on FastAPI + SQLite with Markdown/PDF export and a real consent & ethics layer — not just a disclaimer.
Defensible AI, on your terms. One of several local-first projects I'm building. 👇 https://youtu.be/ZZ1sFozT3fM
