Blocks
Agentic RAG
Advanced retrieval with iterative refinement and multi-source search
Agentic RAG goes beyond simple retrieval by iteratively refining queries, searching multiple sources, evaluating result quality, and re-ranking results. Supports dense, sparse, and hybrid retrieval strategies.
Block Preview
Agentic RAG
Advanced retrieval with iterative refinement and multi-source search
Usage
- Add the block to your workflow and connect it to the upstream step.
- Configure any required credentials or tokens in the inputs.
- Fill in required inputs and optional parameters for the run.
- Run a test execution, inspect outputs, and iterate before deploying.
- Deploy the agentic_rag block with monitoring enabled in production.
Inputs (UI)
Query
long-inputPlaceholder: Enter the question or search query...
Layout: full
Knowledge Source IDs
short-inputPlaceholder: Comma-separated source IDs or use knowledge-source-input
Layout: full
Retrieval Strategy
dropdownLayout: half
Options: Hybrid (Recommended), Dense (Embedding), Sparse (Keyword)
Max Results
sliderLayout: half
Enable Re-ranking
switchLayout: half
Iterative Refinement
switchLayout: half
Max Iterations
sliderLayout: half
Condition: iterativeRefinement = true
Inputs (API)
query
stringRequired
sourceIds
stringRequired
strategy
stringOptional
maxResults
numberOptional
reranking
booleanOptional
iterativeRefinement
booleanOptional
maxIterations
numberOptional
Outputs
Primary response type:
json