UFDR AI is an advanced AI-powered forensic intelligence system designed to revolutionize digital investigations by automating the analysis of Unified Forensic Data Reports (UFDRs). Built for law enforcement, intelligence agencies, and forensic laboratories, it transforms unstructured UFDR data into actionable, explainable, and court-ready insights — reducing analysis time from days to minutes.
- Multi-Format Ingestion: Supports UFDRs from Cellebrite, Magnet AXIOM, Oxygen, XRY, and custom exports (JSON, XML, SQLite, PCAP, Text)
- OCR & Artifact Extraction: Uses Tesseract to extract text from screenshots and image attachments
- Canonical Normalization: Converts vendor-specific formats into a unified forensic schema
- Hybrid Search: Combines keyword-based (BM25) and semantic (embedding-based) search for higher precision
- Natural Language Queries: Ask questions like "Show all chats containing cryptocurrency addresses shared with foreign numbers last month."
- Provenance-First RAG: Every answer includes exact file name, line number, and confidence score
- Entity Extraction: Detects phone numbers, IPs, crypto addresses, device IDs, and user references
- Temporal Knowledge Graph: Explore evolving relationships across people, devices, and communication events
- Timeline Visualization: Chronological view of messages, calls, and events
- Interactive Graphs: Relationship mapping using NetworkX and Plotly
- Anomaly Detection: Flags irregular patterns such as sudden message bursts or new device appearances
- Risk Scoring: Assigns explainable suspicion scores to events
- One-Click Report Generation: Exports findings to PDF or CSV with metadata and confidence levels