CaseCite
An open-source, self-hosted alternative to Harvey. Law firms install it on their own servers, upload their documents, and get research, drafting and contract review cited to the source.
The problem
Law firms want legal AI, but most tools ask them to send privileged client documents to someone else's cloud, and to pay for it per seat.
A legal answer is also only useful if a lawyer can check it. An answer without a citation is a liability.
What I built
CaseCite is a legal RAG platform that a firm runs on its own servers. Documents, search indexes and keys stay in the firm's environment, and AI calls go to the model provider the firm chooses, on its own keys.
Eight tools sit behind one login: research with inline citations, document import from the systems firms already use, clause-by-clause contract comparison with redlines, long-document drafting, and case law tools built on CourtListener data.
What it does
Cited research
Semantic search with BM25 rescoring. Every answer cites the passage it came from.
Document import
From Google Drive, OneDrive, Box, Dropbox, Filevine and Clio.
Contract review
Compare two to four agreements clause by clause and export the redlines.
Drafting
Long documents drafted from the firm's own material.
Case law
Case view, authority maps and judge intel, from CourtListener.
Firm-grade security
MFA, Azure AD single sign-on, AES-256-GCM encryption and a tamper-evident audit log.
By the numbers
Under the hood
- RAG
- FastAPI
- React
- Postgres
- Redis
- ChromaDB
- Docker Compose
- OpenAI, Anthropic or Gemini on the firm's keys
Want a system like this working for your business?
I design, implement, and optimize agentic systems that generate revenue, automate operations, and scale businesses. Tell me what you're working on.









