Google Cloud’s Always-On Memory Agent Replaces RAG and Embeddings With Continuous LLM Consolidation on Gemini 3.1 Flash-Lite
Google just shipped a memory system that kills RAG. No embeddings. No vector DB. Just continuous LLM consolidation on Gemini 3.1 Flash-Lite.
Why it matters
Google Cloud is shipping a production-ready alternative to RAG/embeddings architecture using stateful LLM agents with persistent SQLite memory. This represents a meaningful shift in how builders approach context management and could reshape RAG adoption patterns.
The key facts
6 to knowAlways-On Memory Agent reference implementation ships in Google Cloud generative-ai repository
Built on Google ADK and Gemini 3.1 Flash-Lite
Eliminates vector database and embeddings layer
Uses orchestrator pattern: Ingest, Consolidate, Query sub-agents
Memory persisted in SQLite with 24/7 continuous processing
Treats memory as running process rather than static retrieval
Go to the source
MarkTechPostmarktechpost.com
Publisher excerpt: Google Cloud's generative-ai repository ships the Always-On Memory Agent, a reference implementation that treats memory as a running process. Built on Google ADK and Gemini 3.1 Flash-Lite, it uses no vector database and no embeddings. Instead, an orchestrator routes to Ingest, Consolidate, and…