Lex-Contrast is a futuristic legal intelligence platform that uses a "Swarm" of AI Agents to instantly audit and compare complex contracts.
Instead of one single AI, it deploys a team of specialized agents—like a Liability Steward, IP Sentinel, and Termination Warden—who "handshake" and collaborate to find hidden conflicts, risky clauses, and inconsistencies between two documents.
In short: You upload two files, watch the AI agents swarm together in real-time to analyze them, and get a high-fidelity report showing exactly where you're at risk and how to fix it.
| Infrastructure Initialization | Neural Ingestion | Neural Handshake |
|---|---|---|
![]() |
![]() |
![]() |
| Active Swarm Monitoring | Strategic Synthesis |
|---|---|
![]() |
![]() |
Lex-Contrast operates on a Cortex-Specialist model. Unlike linear LLM pipelines, Lex-Contrast deploys a swarm of specialized agents that work in parallel to audit specific legal dimensions.
graph TD
A[User Input: Document A + B] --> B[Cortex Supervisor]
B --> C[Neural Handshake Initialization]
C --> D{Swarm Coordination}
D --> E[Liability Steward]
D --> F[IP Sentinel]
D --> G[Termination Warden]
E --> H[Conflict Detection]
F --> H
G --> H
H --> I[Strategic Synthesis]
I --> J[Final Audit Report]
- Cyber-Neumorphic UI: A premium, futuristic interface using glassmorphic panels, inset/outset neumorphic shadows, and reactive micro-animations.
- Three.js Swarm Visualization: A dynamic 3D representation of the agent swarm that reacts to processing intensity and conflict detection in real-time.
- Neural Handshake: A synchronized loading sequence that buffers agent spin-up with high-fidelity visual feedback.
- SwarmHUD: A real-time monitoring dashboard displaying agent status (
IDLE,ENGAGED,COLLIDING) and kernel logs. - Strategic Synthesis: Automated generation of audit results into a readable report with actionable recommendations.
- Framework: Next.js 15 (App Router)
- Styling: Vanilla CSS with custom Neumorphic & Glassmorphic tokens.
- 3D Engine: Three.js with
@react-three/fiberand@react-three/drei. - Animations: CSS Keyframes +
framer-motionfor fluid state transitions. - Icons: Custom Lucide-React integration.
- Node.js 18+
- NPM / Bun / PNPM
- Clone the repository:
git clone https://github.com/abhinavv27/LESt.git cd web - Install dependencies:
npm install
- Run the development server:
npm run dev
- Access the dashboard at
http://localhost:3000.
You can perform a manual audit using the provided mock documents:
- Upload doc_a.txt as the Baseline Source.
- Upload doc_b.txt as the Contrast Source.
- Execute the RUN_NEURAL_AUDIT sequence to witness the swarm in action.
The specialized agents in the Lex-Contrast swarm are derived from frontier legal-tuned models.
- Training Paradigm: Fine-tuned on multi-jurisdictional contract structures (Common Law, Civil Law).
- Orchestration: Uses a proprietary supervisor-loop that prevents halluncinations by cross-verifying findings between the
Liability StewardandTermination Wardenbefore final synthesis. - Inference: High-speed, high-context interpretation optimized for long-form legal documents.
Lex-Contrast reduces the "First Pass" review time by up to 85%, acting as a force multiplier for legal teams during high-volume M&A activity or Master Service Agreement (MSA) negotiations.
Distributed under the MIT License. See LICENSE for more information.




