Factagora

About Factagora

Factagora is a verification layer you call before your model's answer reaches a user. Three core endpoints sit behind one bearer-authenticated key with an OpenAPI spec: GET /fact-search returns ranked, cited results from verified sources for a query; POST /fact-checker takes a claim and returns a verdict with a confidence score plus the exact source documents behind it; POST /evidence-finder surfaces both supporting and opposing evidence ranked by strength, which is the endpoint that matters when a question is genuinely contested rather than simply true or false. Additional APIs cover deep research, timeseries, causality graphs and a fingerprint family for embedding, detecting and reporting on content provenance. The architectural claim is that it is additive: Factagora sits between your application and whatever LLM, vector database or RAG stack you already run, so adoption is one call rather than a migration or a retrain. The vendor's structure is a temporal knowledge graph of FactBlocks rather than a vector store, with source credibility scored explicitly, and its published case study is a top-five Korean law firm that restructured over 350,000 documents into it. A free playground with welcome credits lets you test before wiring anything in. Adjacent products in the same family include a DeepVerify browser extension and a DeepStamp provenance mark, and a playground lets you exercise the endpoints in the browser before writing any integration code.

Key Features

Tags