Current stack
- Cloud Run serves one FastAPI runtime
- Cloud SQL runs Postgres and pgvector
- Vertex AI supplies Gemini generation
gemini-embedding-001supplies text embeddings
Keep changing facts straight.
Brain keeps a sourced history of the messages, documents, events, and records behind your product. For each request, it returns the material that applies, where it came from, and what remains uncertain.
How Brain works.
01
Keep the record
Save source material with the time it was recorded and where it came from.
source → Record02
Track the change
Derive relationships and versions without erasing the source.
Record → Memory03
Leave decisions with your app
Keep application-approved Objects separate from derived Memories.
Memory ≠ Object04
Return the evidence
Return the records a request may use, with citations, gaps, conflicts, and cost.
query → ContextBuilt to be inspected.
Cloud Run serves one API. Cloud SQL stores Records, Memories, Objects, and embeddings. Vertex AI derives and ranks. Sources, time, and application decisions remain visible in the result.
What is not included
BigQuery can send structured rows through Records, but the managed BigQuery connector is not shipped. Model and embedding upgrades are tested before they replace the current runtime.
Source, finding, decision, result.
A
Source
Records retain content, stable customer IDs, metadata, and source time.
what enteredB
Memory
Memories retain versions and citations. Objects carry facts your application accepts.
what changedC
Context
Context selects the evidence this request may use and names what remains unknown.
what applies nowYour warehouse and application remain authoritative for operational data.
Start with one Record.
The quickstart shows the source, the request, and the returned evidence.