Demo store · Knowledge graph guide
The FactoGraph guide: a knowledge store about knowledge stores
A public, read-only store that explains FactoGraph to people and AI agents. It is the store the FactoGraph MCP server reads when no other store is named, so asking it is also the quickest way to see how a cited answer looks.
Questions to try
- How is FactoGraph different from a vector database?
- Does using a FactoGraph store need an account or an API key?
- Can an AI agent store facts in FactoGraph without my approval?
- What does a FactoGraph store reply when it cannot answer a question?
- How can a team of agents share one FactoGraph store?
What the store covers
How it compares
FactoGraph against chat memory built into one assistant, plain notes files, vector databases, retrieval over raw documents and self-hosted memory.
What makes it different
Human approval of every fact, contradiction checks before storage, relation layers called spaces, derived facts, and answers cited fact by fact.
Scenarios
Research, project, compliance, support, coding and multi-agent teams that work from one shared store.
Questions agents ask
Price, accounts and API keys, which models work, approval before storing, what happens when the store cannot answer, and correcting facts.
Connecting
Reaching a store by its link, over MCP or by plain HTTP, and the difference between a private link and a public one.
Where the facts come from
- FactoGraph: a guide for AI agents
- Why an AI agent benefits from FactoGraph
- FactoGraph for agents: comparisons, scenarios and a quickstart
The guide describes FactoGraph itself and is kept by the FactoGraph team. Its facts are the same ones the guide page for AI agents is rendered from.