Context recall
How one call turns a question into prompt-ready memory: hybrid retrieval, entity expansion, and the weakEvidence abstention signal.
Context recall is the read primitive everything else exists for: one call, one prompt-ready block. See the API reference for parameters; this page explains the behavior.
What happens on a recall
- Hybrid retrieval (BM25 + vector) ranks documents for
q- each hit reports whichsourcesmatched. - The top entities are expanded with their current facts and recent timeline.
- Everything is fused into
contextMd, ordered for prompt injection.
Abstention beats hallucination
If the best hit scores under 0.5, the response sets weakEvidence: true. Your agent should treat that as "the brain does not know" - caveat the answer or proceed without memory, never present thin matches as established fact. Memory layers that always return something teach agents to invent provenance.
Recall discipline for agents
- Recall at task start, not mid-stream - one good
contextMdinjection beats five micro-queries. - Scope with
pathPrefixwhen the task has a lane (/workspace/teams/platform/). - Use
includeBodiesonly when the caller won't make follow-updocreads (single-shot pipelines).
How Unison works
Architecture in one page: agents ingest conversations and documents, an async pipeline extracts entities and bitemporal facts into a knowledge graph, and any agent recalls the relevant slice as markdown.
Knowledge graph
Entities, bitemporal facts, and links: how Unison stores what is true, when it became true, and how knowledge connects.