FactBlock
The smallest unit of trusted knowledge
Overview
FactBlock is the smallest unit of trusted knowledge. It extracts key facts from unstructured text, structures them with metadata and sources, and makes them reusable for AI, automation, and decision-making.
Structure
Each FactBlock consists of the following components.
- Claim — A verifiable factual claim
- Source — Source information for the claim
- Timestamp — When the fact was valid
- Metadata — Additional context information
How it Works
- Input text, documents, or URLs.
- AI extracts and structures key facts.
- Sources and timestamps are automatically linked.
- FactBlocks are stored in collections for verification.
Use Cases
- Break down news article claims into verifiable units
- Extract and track key figures from financial statements
- Store research paper conclusions in structured format
- Convert legal document clauses into searchable format