RAG Retrieval & Grounding Diagnostics
Assess user-reviewed retrieval relevance, answer support and query coverage using consistent evaluation counts.
← Explore the full tool libraryEnter your scenario
The form starts with hypothetical examples. Replace them with your own measurements or assumptions. Inputs and exports stay in this browser.
What decision does this help answer?
Is a retrieval problem causing unsupported answers, or is the answer generator ignoring evidence?
Calculation and methodology
Compute document-level precision from relevant retrieved passages divided by retrieved passages, query coverage from queries with sufficient relevant passages divided by all queries, and judged answer support from supported answers divided by checked answers. Each dimension is displayed separately; no single composite score hides whether the retrieval or generation stage needs work.
Limitations and responsible interpretation
All labels and counts are user-supplied, not scraped or judged by Ninth Post. Relevance is not the same as evidence sufficiency. Use a stated query sample, annotated passages, a consistent claim-support rubric and additional per-query metrics before acting on the result.
Research and supporting guidance
This planning tool complements Ninth Post's related AI product analysis. Consult the underlying technical guidance or official framework where applicable, and see our editorial methods. The tool does not independently validate submitted assumptions, current vendor prices or legal obligations.