CaseCite: An Open-Source Self-Hosted Legal RAG Platform
A private, infrastructure-controlled alternative to enterprise legal AI research tools.
1Open-Source Legal Research Alternative
CaseCite is an open-source, self-hosted platform designed for legal document analysis and case research. It allows law firms to maintain full control over their sensitive data by hosting the entire stack on their own servers. Unlike proprietary cloud-based solutions, it ensures that no document data is sent to external maintainers.
2Infrastructure and Data Sovereignty
The platform prioritizes data privacy by keeping documents, vector indices, and chat history within the user's infrastructure. It supports a Bring Your Own Key (BYOK) model, where users configure their own AI provider API keys. This architecture ensures that firms retain complete custody of their information throughout the research process.
3Technical Stack and Deployment
CaseCite is built using a containerized architecture, requiring Docker and Docker Compose for deployment. The stack includes a FastAPI backend, a PostgreSQL database, and Redis for session management. It provides flexible configuration options, allowing users to deploy behind existing reverse proxies like Traefik or use a bundled production stack.
4AI Integration and RAG Pipeline
The system utilizes a Retrieval-Augmented Generation (RAG) pipeline to process legal documents and case law. It integrates with CourtListener for case-law data and supports various embedding providers. Users can maintain document privacy by hosting their own OpenAI-compatible servers for both chat and embedding models, keeping all text within their network.
5Security and Authentication Features
Security is managed through PBKDF2-SHA256 password hashing, optional TOTP MFA, and JWT-based authentication. The application implements AES-256-GCM encryption for uploaded documents and API keys at the application layer. Additionally, it features HMAC-signed audit logs and per-IP rate limiting to protect against unauthorized access and ensure consistent system integrity.
6Practice Profile and Context
CaseCite features a dynamic practice profile system that allows lawyers to define their specific practice area. This profile acts as standing context for every review, redline, and draft generated by the AI. Users retain full ownership of these profiles, which are saved via settings and applied across all legal matters.
7Connector and Integration Capabilities
The platform includes connectors for various document management systems, such as Clio, Filevine, iManage, and NetDocuments. While some connectors are implemented against vendor APIs, they require validation in live environments. Microsoft and Google integrations are also supported, requesting read-only scopes to ensure minimal access to the firm's existing document repositories.
8Licensing and Community Contribution
Released under the Apache License 2.0, CaseCite is a source-available project that encourages community contributions. Developers can submit pull requests following the Developer Certificate of Origin sign-off process. Commercial support and private installation services are available directly from the maintainer, providing professional assistance for firms requiring specialized deployment or maintenance.
9Research Aid Limitations
CaseCite is positioned strictly as a research aid and not a substitute for professional legal judgment. Every AI-generated output carries a mandatory notice to verify information against source documents and current law. It does not hold security certifications like SOC 2 or HIPAA, placing the responsibility of host-level encryption on users.
Original article: Hacker News — CaseCite: An Open-Source Self-Hosted Legal RAG Platform