GETTING STARTED
Connect your first Agent.
Give your knowledge a space, then let your Agent model and maintain it from your sources.
Create a knowledge space
Register and sign in, name the space and describe the problem it should solve. No manual ontology design is required.
Issue an Agent credential
Create a maintenance connection under Agents & runs. The credential is shown once; save it in your Agent’s secret configuration.
Connect MCP and supply sources
Copy the MCP endpoint and connection settings from the console. Give selected sources to the Agent, and ask it to read context, define the model, save sources and submit knowledge changes.
Review what persists
Explore the model, knowledge graph, evolution and source evidence from the overview. Verify what the Agent actually produced.
Your first instruction to the Agent
Read the current space context. Build a domain model from my materials, save traceable source snapshots, and write evidence-backed objects and relations. Link the changes to a maintenance run and report which changes were applied and which need confirmation.