WILDSCAN
Autonomous six-agent system that detects wildlife trafficking on regional marketplaces via automated scraping, multilingual code-word matching, vision species ID, and risk scoring.
Overall Winner - Unicorn Mafia x Techbible Hack Night 2026
Built in 90 minutes; awarded the overall prize at the London hack night (Web MCP Agents track), March 2026.
Challenge
Wildlife trafficking moves through fragmented regional marketplaces in dozens of languages, faster than human investigators can monitor. Existing tooling stops at translation.
Approach
Built a six-agent LangGraph pipeline running on Bright Data Web MCP, paired with a 500-term coded-language lexicon across 8 languages and deterministic risk scoring against CITES/IUCN classifications.
What it does
6-Agent LangGraph Pipeline
Scanner, Triage, Linguist, Image Analyst, Species Classifier, and Risk Scorer running fully autonomously.
Multilingual Lexicon
500 coded trafficking terms across 8 languages with fuzzy matching for obfuscated listings.
Deterministic Risk Scoring
0-100 scores from 8 weighted signals, correlated against 6,000 historical seizure records.
Globe Visualization
Real-time Mapbox globe plotting detections geographically with on-demand intelligence briefs.
Gallery
