Factlens

FactLens is a political intelligence and fact-verification platform built for Indian news. In an era of rapid misinformation, FactLens automates the process of verifying political claims by grounding every verdict in real, retrieved evidence — never in model assumptions or training knowledge.

The platform ingests news from Indian sources including RSS feeds, SEBI orders, NSE/BSE data, and Indian Kanoon judgements. It extracts named entities, typed relationships, and factual claims using NLP, and stores structured intelligence across a purpose-built multi-database architecture — PostgreSQL with TimescaleDB for time-series article storage, Qdrant for semantic vector search, and Neo4j for entity relationship graphs.

When a user submits a claim, FactLens performs hybrid retrieval — combining semantic similarity search with structured entity queries and graph traversal — then passes the assembled evidence to an LLM to produce a structured verdict with citations to source articles.

FactLens is designed for journalists, researchers, and citizens who need fast, traceable answers to political questions — built on the principle that every conclusion must be explainable and every source must be cited.