Legal document AI

Habeas Context

A private workspace for legal drafting, research, live notes, and document review where every client file stays organized in a Git-backed record.

Git-as-DB

Client documents live in versioned directories for reviewable history and clear provenance.

Local auth

User accounts and roles are scoped to each client workspace rather than a shared global store.

Focused AI

Chat, citations, drafts, live notes, and file workflows are built around legal work product.

Workflow

From case material to drafted work product

01

Collect

Bring client files, transcripts, research notes, and supporting material into a structured workspace.

02

Discuss

Use context-aware chat and concern threads to reason over files without losing the surrounding matter.

03

Draft

Move from notes to citations and editable drafts while preserving a record of how documents changed.

Principles

Built for accountable legal work

Auditability first

Git history is treated as a product primitive, so file changes can be inspected instead of hidden behind opaque storage.

Client-level boundaries

Configuration, users, and files are organized around a client path, keeping each matter's operational context explicit.

Practical AI surfaces

The interface centers on repeated legal workflows: documents, drafts, conversations, history, library, and live notes.

Stack

Python backend. React frontend. Plain files underneath.

Backend
Python server routes for auth, files, chat, drafts, history, settings, library, and live notes.
Frontend
Vite and React application with collaborative editor and legal document workflows.
Data model
Client-specific directories hold documents, user records, and the Git-backed working record.