Find a known entity
Load an entity by its identifier or use an indexed field for an exact lookup.
03 / Retrieval
Get the knowledge your next action needs.
Find current facts, follow relationships, and inspect past changes. Give your agent a focused view of the record for the task at hand.
In practice
Ask about the project. Follow the context.
Illustrative record. Current facts, relationships, and historical values can be queried for different questions.
- Current status
- Active
- Project owner
- person:alice
- Related team
- team:platform
- History
- Previous status: planned
Built into the workflow
Several ways to reach the right record.
Search and explore
Match text in current entities and facts, then traverse typed relationships to discover connected records.
Look back when needed
Inspect the changelog or retrieve an entity at a point in valid time to understand what was recorded as true then.
Take a readable view with you.
Export current knowledge as JSON or Markdown for inspection and documentation. Exports are derived views; submit observations to update the stored record.
Connect retrieval to your agentCOMMON QUESTIONS
Retrieval, answered.
Is retrieval semantic or vector search?
Neither. Search is text matching over current entities and facts. Retrieval also offers exact lookup by identifier or indexed field and traversal of typed relations. Use ConceptMem alongside a document retrieval system rather than in place of one.
How do I ask what was true at a past date?
Facts are bitemporal. A point-in-time read resolves an entity against valid time, meaning when the value was true of the world, falling back to transaction time for records written before valid time was captured.
What is the difference between the changelog and history?
History is the set of superseded and retracted facts still attached to an entity. The changelog is the immutable record of mutations, with the provenance of each one. History answers what the value used to be; the changelog answers how it changed.
Can I take memory out of the system?
Yes. Current knowledge can be exported as JSON or Markdown. Those exports are derived views: editing one changes nothing, and updates are made by submitting an observation.
Make your next update count
Start with one useful memory.
Create a workspace, define a type, and see your first observation become a traceable record.