Whether init() has completed successfully.
Number of configured corpus directories.
Number of loaded markdown chunks in the in-memory corpus.
Number of extracted topic entries used by the classifier.
Number of unique source files represented in the loaded corpus.
Counts for the bundled platform knowledge corpus currently loaded in memory.
Whether retrieval is vector-backed or keyword-only.
Whether corpus embeddings are active, missing credentials, or failed during init.
Embedding dimension for the active vector index, or 0 when inactive.
Whether graph expansion is enabled in config.
Whether deep research is enabled in config.
Runtime truth for graph expansion.
Runtime truth for reranking.
Runtime truth for deep research.
Lightweight observability snapshot for router startup logs and host health checks.
Returned by
router.getCorpusStats()after or before initialization.