Retrieval
Retrieval systems that stay grounded
Combine file search, embeddings, and memory notes to answer with citations and provenance.
Use retrieval, vector search, and memory notes to ground responses in trusted data.
- Attach source metadata to every chunk.
- Add a reranker for high precision.
- Return citations with each answer.
- Expire memory notes with TTL.
Chunking strategy
Use overlap and metadata for consistent retrieval.
Reranking
Re-score results for relevance and reduce noise.
Citations
Return sources alongside every answer.
Memory notes
Persist safe notes for long-lived workflows.
Retrieval + memory guides
Curated recipes, playbooks, and walkthroughs for this topic area.
Connector integrations guide
Connect agents to Gmail, Outlook, CRM systems, and calendar apps.
Context engineering with memory notes
Store long term memory notes and state for multi-session workflows.
Multi-tool orchestration with the ADK
Plan, call tools, and summarize into a single runbook.
RAG on PDFs with file search
Chunk documents, search with citations, and answer precisely.
Summarizing long documents
Chunk, summarize, and merge outputs with quality checks.