Forensic AI — Local Offline Conversational Forensics

Overview

This repository is an initial scaffold for a fully-offline, deterministic, auditable conversational forensics system.

Goals

• Single Python computational core shared by CLI and GUI (via local REST API)
• Deterministic processing with run/seed reproducibility
• Structured JSON logging and immutable run manifests
• DuckDB/SQLite for tabular storage and optional Neo4j/Redis components (local-only)
• Extendable plugin system and offline packaging

This scaffold contains a minimal FastAPI server exposing the core endpoints required by the spec, a CLI, a sample config, and a small test.

Next steps

• Implement ingestion, preprocessing, storage adapters (DuckDB + graph sync)
• Add Ray and PyTorch-based analysis modules
• Scaffold Electron+React GUI and packaging
• Add unit/integration tests and load testing harness

See config/config.yaml for configuration options and pyproject.toml for dependencies.