Association-DataRing

Association-DataRing / Louis

Public

Plateforme IA juridique souveraine, auto-hébergeable et open-source. Bring Your Own Key, agrégateur de modèles FR/EU/US, RAG pgvector, orchestration multi-agents.

louis.data-ring.net agpl ai-sdk anthropic byok french
10
1
85% credibility
Found May 24, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Louis is an open-source AI platform built specifically for legal professionals in France. It lets lawyers create a private AI assistant that coordinates multiple specialized agents — for legal research, document drafting, review, and citation — all while keeping full control over their data and AI infrastructure. Users connect their own AI provider accounts, upload their legal documents, and can generate or edit contracts and formal letters. Everything runs on the user's own server, with no data sent to external services. The platform includes integrations with French legal databases (Légifrance, Pappers), document management with semantic search, a library of reusable legal workflow templates, multi-user access with role-based permissions, and detailed cost tracking. It's designed as a sovereign alternative to commercial AI tools, respecting legal professional secrecy by keeping sensitive client information entirely under the lawyer's control.

How It Works

1
⚖️ You hear about a new kind of AI tool for lawyers

A colleague tells you about Louis — an AI assistant that stays under your control, unlike cloud services that send your clients' documents elsewhere.

2
🔑 You connect your own AI service

In the settings, you add your own AI provider account — whether Mistral, Anthropic, or another. Your keys, your account, your control.

3
📂 You upload your legal documents

You drag and drop PDFs and Word files into Louis. It reads them, extracts the text, and remembers everything so you can ask questions about your own files later.

4
💬 You start a conversation about a legal matter

You type your question in the chat — maybe about a contract clause or a recent court decision. You can attach documents or ask it to search French legal databases.

5
You choose how your AI team works
⚡
Quick chat mode

Ask a question, get an answer — fast and simple, with your documents and legal databases at your fingertips.

🔬
Multi-agent mode

Watch a team of agents work together — one researches, one drafts, one reviews — and see each step in the audit trail.

6
📄 You generate or edit a legal document

You ask Louis to draft a formal letter or contract. It creates a Word document, or suggests tracked changes to an existing one that you can accept or reject.

✅ Everything stays under your control

Your data never leaves your own infrastructure. You have a full audit trail of every AI response. Your clients' information is protected — on your terms.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is Louis?

Louis is a self-hostable legal AI platform built in TypeScript with Next.js. Instead of another chatbot, it functions as an orchestration layer where multiple specialized AI agents (research, drafting, reviewer, citator) work together under a master coordinator to produce traceable, auditable legal work. You bring your own API keys for Mistral, Anthropic, OpenAI, Albert, Scaleway, or OVH -- nothing routes through DataRing. The platform includes RAG-powered document search via pgvector, legal connectors to Légifrance through the PISTE API and Pappers, and generates formatted DOCX/PDF documents with tracked edits. Multi-user support with role-based access, per-model cost tracking, and an append-only audit log complete the offering.

Why is it gaining traction?

The privacy angle is the hook. Legal professionals cannot send client documents to US-based SaaS vendors without triggering uncomfortable questions about Cloud Act and FISA 702 extraterritoriality. Louis lets firms run the same AI capabilities on their own infrastructure, with their own keys, under their own control. European providers are prioritized in the interface. The architecture is genuinely multi-agent -- not just a single LLM call -- which means each response can include an audit trail of which agent did what. The MCP-native design means you can wire in custom legal tools without waiting for upstream features.

Who should use this?

French and EU law firms, in-house legal teams, and legal tech builders who need sovereign AI infrastructure. It is particularly relevant for due diligence workflows, contract analysis, and any practice where attorney-client privilege makes cloud AI risky. Solo practitioners and small firms who want AI assistance without surrendering data to US vendors will find the BYOK model appealing. Developers building legaltech products could use it as a foundation since the code is AGPL-3.0 and auditable.

Verdict

At 10 stars and v0.1, Louis is early-stage but the architecture is solid and the security model is thoughtful. With a credibility score of 0.8500000238418579%, it is credible enough to evaluate but not production-ready for client work without your own testing. The documentation is thorough, the Docker Compose setup is one-command, and the feature set (RAG, multi-agent, document generation, audit trail) covers real legal workflows. Watch for v0.2 multi-agent expansion before committing to a production deployment.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.