Cloning Legal AI: The Mike Open-Source Project
Indie Hacker Newsgo watch the original →
the gist
A former Big Law associate built an open-source legal AI platform called Mike that replicates core enterprise features like bulk document review and citation-backed chat, challenging the high-cost subscription models of incumbents like Harvey.
The Core Functionality
Mike is an open-source legal AI platform designed to perform the same document analysis tasks as enterprise incumbents without the per-seat licensing fees. The application provides a chat interface for document interrogation, a project-based workspace for managing legal matters, and a tabular review engine that extracts data across hundreds of documents in parallel. Every extraction includes verbatim citations back to the source document to prevent hallucinations. The system also includes a workflow engine that allows users to save and share prompt templates for common legal tasks such as change-of-control, NDA, and credit agreement reviews.
Technical Architecture
The project is built on a standard, accessible stack that allows for self-hosting in approximately 10 minutes. The frontend is built with Next.js, while the backend utilizes Express. The application relies on Supabase for authentication and PostgreSQL, with Cloudflare R2 serving as the S3-compatible storage for documents. LibreOffice is used to convert Word documents into PDF format for model ingestion. The system is model-agnostic, allowing users to configure API keys for Claude, Gemini, or OpenAI directly. Because the platform uses a bring-your-own-key model, users pay model providers directly for token usage without additional platform markups or annual minimums.
Market Impact
The project highlights that the primary value proposition of enterprise legal AI tools is often the interface and vendor trust rather than proprietary model training or custom embeddings. By providing a functional, self-hosted alternative, the project forces a shift in renewal conversations for law firms. While it lacks enterprise-grade features like Westlaw integration or audit logs, it provides a viable option for solo practitioners, boutique firms, and in-house teams who previously could not justify the high costs of enterprise legal AI contracts.