

OSINT-D2 v2 — Open source CLI tool with autonomous AI agent, trust anchors, and breach intelligence
Built this after getting tired of juggling 5 different tools. v2 adds:
Agent Mode — Give it a username or email, an LLM autonomously investigates: scans 19+ platforms, pivots on findings, profiles the target across 6 dimensions (identity, geo-temporal, psychological OCEAN, technical, ideology, attack surface).
Trust Anchors — Declare known accounts (Instagram, TikTok, Facebook, Telegram, Email) as ground truth. The tool verifies everything else against them and discards false positives. Works surprisingly well.
Breach Intelligence — HIBP integration. Maps every email to known breaches with exposed data classes.
Proxy Support — Residential & datacenter via ScrapingAnt. Geo-targeted. Your own keys, your own scans. Nothing phones home.
PDF Dossiers — Generates a classified-style PDF with cover page, identity cards, confidence meters, breach tables. Looks like it came from a three-letter agency.
Built in Python. Uses Typer + Rich for CLI, WeasyPrint for PDFs, Jinja2 templates, OpenAI-compatible API for the agent.
Free. MIT. Local-first. Built solo over the past months.
https://github.com/Doble-2/osint-d2
Happy to answer questions.