Contribution to a street-level OSINT pipeline
Pi Geosint Agent
Contributed to a browser-controlled pipeline for discovering, capturing, and storing street-level imagery on a Raspberry Pi.
Current state
Open-source contribution to a local web tool with persistent evidence storage and a documented Docker smoke test.
Case study
Problem
Collecting street-level imagery evidence requires coordinating location parsing, multiple sources, capture, and durable metadata.
Approach
I contributed to a staged pipeline that resolves a place, discovers imagery, recommends captures, and stores evidence while deterministic tools retain control.
Engineering
The shared codebase uses TypeScript for the web UI and pipeline, Docker for the runtime, and optional model providers for bounded decisions.