Building BitLaw
An AI-powered legal platform - a "Second Brain" for lawyers.
I'm building BitLaw, an AI-powered legal platform designed to be a "Second Brain" for lawyers. It combines document management, AI chat, contract analysis, and legal research into one integrated system.
Status: In development | Time so far: 20 days | Deadline: Mid-February
The Problem
Legal work involves massive amounts of documents, case law, contracts, and deadlines. Lawyers spend hours searching through files, cross-referencing information, and drafting documents that follow specific patterns.
Current tools are fragmented. Document management is separate from research. Research is separate from drafting. AI assistants don't know the context of what you're working on.
What BitLaw Does
Jurisdiction-Aware Knowledge — When an organization creates an account, they specify their jurisdiction in LATAM. A shared wiki serves as the trusted legal knowledge base for that jurisdiction, curated and available to all users. The AI prioritizes sources in this order: jurisdiction wiki, then organization documents, then internet search.
Context Engine — The AI automatically knows what documents belong to the matter you're working on. Ask "What are the key deadlines?" and it answers using your actual case documents.
Document Editor — The AI can create documents from scratch. Edit them in a rich text editor with full version history. Share via link for viewing and comments. Export to DOCX matching your firm's format.
Tabular Review (in development) — Upload 50 vendor contracts and the AI extracts termination dates, liability caps, renewal terms into a spreadsheet view.
Document Management — Upload, view, and organize documents by matter.
Matters & Clients — Case management and client relationship tracking.
Research — Legal research interface with AI assistance.
Why I'm Building This
Legal AI tools either try to do everything poorly or do one thing well but don't integrate with the rest of the workflow.
BitLaw is designed around how lawyers actually work: they have a matter, that matter has documents, and they need to chat, draft, and research within that context.
The AI should know what you're working on without you having to explain it every time.
Under the hood, we use Helix DB for its graph-vector functionality. Legal documents constantly cross-reference each other—contracts cite clauses, cases cite precedents, memos reference exhibits. A graph-native database captures these relationships in a way traditional vector stores can't.