fit-rag

name	fit-rag
description	Query the knowledge indexes: search by meaning, answer relationship questions with triple patterns, or list graph subjects by type. Use when you need ranked semantic results or graph relationships without standing up a vector database or SPARQL endpoint.

Query the Knowledge Indexes

fit-rag is the read surface over the indexes the build pipeline produces. One binary, three subcommands: search ranks a vector index by meaning, query matches a triple pattern against a graph index, and subjects enumerates graph subjects by type. Build the indexes first; this queries them.

When to Use

• Find content related to a phrase by meaning — fit-rag search 'career progression'
• Answer "which X relates to Y" without join logic — fit-rag query "?" rdf:type schema:Person
• Enumerate the entities of a type before forming a query — fit-rag subjects schema:Person

Usage

# Semantic search — one result per line, identifier<TAB>score, top 10 by
# descending similarity. Calls the embedding service to vectorize the query.
fit-rag search 'career progression'

# Triple-pattern query — exactly three arguments (subject predicate object).
# Put `?` in the position to resolve; prints one matching identifier per line.
fit-rag query "?" rdf:type schema:Person

# Subject listing — one subject<TAB>type per line; pass a type to filter.
fit-rag subjects
fit-rag subjects schema:Person

search reads the vector index and needs the embedding service; query and subjects read the graph index and run offline. Build and populate the indexes first — see the guides below.

Documentation

• Search Semantically — Find related content by meaning with ranked results from a vector index, no vector database required.
• Query a Knowledge Graph — Answer relationship questions from an RDF graph index with triple-pattern queries and type-filtered subject listings.
• Give Agents Typed, Retrievable Knowledge — The full workflow for building the graph and vector indexes from HTML knowledge sources, then querying them.