Upload contracts. Chat with them in plain English. Extract clauses, assess risk, track obligations, and generate executive reports — all powered by AI.
Document Analysis
Acme Corp – TechStart Acquisition Agreement
3
Critical
7
High
12
Medium
24
Low
Extracted Clauses
Knowledge Graph
6 entities · 7 relations
End-to-end pipeline from PDF upload to actionable insights in six stages
PDF stored in MinIO (S3)
4-tier OCR pipeline
pgvector embeddings
Clauses, risk, entities
RAG-powered Q&A
Reports, obligations, deals
Enterprise-grade contract analysis capabilities built from scratch
Ask questions in plain English. RAG-powered Q&A retrieves relevant clauses and generates contextual answers with source citations.
16+ clause types automatically identified and classified: termination, indemnification, IP, non-compete, and more.
One click to translate any legal clause into simple language anyone can understand. No law degree required.
Each clause scored Critical / High / Medium / Low with specific risk factors and AI-generated summaries.
AI extracts obligations, responsible parties, and deadlines from clauses. Track status as pending, completed, or overdue.
AI-generated executive summaries with risk overview, key clauses, and actionable recommendations. Export as PDF.
Automatically extracts key dates — effective, expiration, renewal, payment — and displays them on an interactive timeline.
Group related contracts into deals for aggregate risk analysis, batch uploads, and portfolio-level insights.
Entity extraction and relationship mapping. Visualize parties, dates, amounts, and obligations as an interactive graph.
Semantic + keyword search powered by pgvector embeddings. Find relevant clauses using natural language.
In-app PDF viewing with clause navigation. Export analysis as PDF, Excel, Word, CSV, or JSON.
Full theme toggle with persistent preferences. Professional interface that adapts to your working environment.
AI identifies and classifies clause types common in M&A contracts
Modern, production-grade architecture with full Docker deployment
Retrieval-augmented generation combines vector search with LLM inference for contextual Q&A
Celery workers handle OCR, embedding, and AI analysis without blocking the API
Every action logged — uploads, analysis, chat questions, exports — with timestamps and context
Server-side API proxy hides backend IP. Docker Compose with 6 containerized services
┌─────────────────────────────────────┐
│ Next.js Frontend │
│ Dashboard · Chat · Deals · Search │
└──────────────┬──────────────────────┘
│ API Proxy (server-side)
┌──────────────▼──────────────────────┐
│ FastAPI Backend │
│ 8 API routers · RAG · Extraction │
├─────────┬───────────┬───────────────┤
│ Celery │ pgvector │ MinIO │
│ Workers │ Embeddings│ (S3 Storage) │
├─────────┼───────────┼───────────────┤
│ Redis │ PostgreSQL│ Ollama │
│ (Queue) │ (Data) │ (LLM AI) │
└─────────┴───────────┴───────────────┘Upload a contract and chat with it. Explore clause extraction, risk assessment, obligation tracking, deal grouping, and knowledge graph visualization.