Open Issues Need Help
View All on GitHub LLM limits tables cover only frontier models, so cheap tiers log 'unknown model' and fall back to guessed limits about 1 month ago
enhancement good first issue
Sparse vectors written at embed time, locked like the dense model about 2 months ago
help wanted
Prefactor: explicit named dense vector in the Qdrant collection about 2 months ago
help wanted
Fixture eval: hybrid vs dense-only retrieval quality on a real book about 2 months ago
help wanted
Hybrid search by default in the CLI (RRF fusion of dense + sparse) about 2 months ago
help wanted