Harvey raised 500M for legal AI, I built the same and open sourced it

CaseCite is a self-hosted legal research platform that allows users to upload their firm's documents and ask questions to conduct legal research. The platform is designed to be run on infrastructure controlled by the user, ensuring data privacy and security. Documents, the search index, and chat history are all stored on the user's own servers.

The AI features in CaseCite transmit user's questions and relevant document text to configured AI providers, with users retaining control over their API keys (BYOK). The platform does not transmit any data to the CaseCite maintainers, and there is no telemetry involved. CaseCite offers eight tools behind one login, all reading from the same document corpus.

Users can specify their practice areas, and CaseCite will generate a profile based on that information, which is then carried along with every review, redline, and draft. Before relying on AI-generated answers, analyses, or drafts, users must verify their accuracy against the source documents and current law. CaseCite is licensed under the Elastic License 2.0 and is not open source (source-available).

Users can deploy CaseCite on their infrastructure by following the provided instructions and setting the required environment variables. The platform does not hold any security certifications and is intended as a research aid, not a substitute for a lawyer's judgment or a citator. Users can get a free CourtListener API token to access case-law features, which are necessary for the product to be fully functional.