Data Pipelines
How evidence moves from upload to a signed, verifiable dossier. (Replace the placeholder with the latest pipeline diagram.)
Stages at a glance
Submit & protect — Users upload chats, screenshots, or receipts. The API normalizes content and immediately tokenizes PII so only masked data flows downstream.
Extract & enrich — OCR and classifiers find entities (wallets, emails, phones), detect language/sentiment, and tag scam type. Related cases are linked.
Store safely — Canonical PII is locked in the vault; case data lands in the structured store + SQL for filters; evidence files go to protected buckets.
Search & triage — Hybrid search (keywords + semantic vectors) lets analysts and LEOs find patterns without exposing identities.
Report & sign — The report generator assembles findings into a dossier, signs it (hash manifest + signatures), and shares via controlled links.
Feedback loop — Outcomes and signals feed back into detectors and playbooks, improving precision over time.
Last updated