Bug Fixes

• cli: surface global output flags in help
• search: preserve generated regex hits and global flags

Full Changelog: v0.12.12...v0.12.13

Bug Fixes

• local: align CoreML reranker and search filters

Full Changelog: v0.12.11...v0.12.12

Bug Fixes

• local: correct CUDA toolkit detection from env

Full Changelog: v0.12.10...v0.12.11

Bug Fixes

• local: stabilize CUDA runtime detection and cache reuse
• parsing: preserve symbol depth and rst update refs

Performance

• parsing: eliminate double parsing and optimize tree-sitter traversal

Full Changelog: v0.12.9...v0.12.10

Bug Fixes

• embedding: clamp gemini batch size

Full Changelog: v0.12.8...v0.12.9

Bug Fixes

• embedding: collapse embedded data blobs before embedding

Full Changelog: v0.12.7...v0.12.8

Bug Fixes

• embedding: adaptive token-aware truncation for context limit errors

Full Changelog: v0.12.6...v0.12.7

Features

• Merge pull request #11 from ineersa/rst-support
• indexing: add RST and Sphinx-aware chunking

Bug Fixes

• Windows DirectML/CPU setup, API token limit parsing, npm binary shim
• indexing: track RST include changes without duplicate chunks

Full Changelog: v0.12.5...v0.12.6

Bug Fixes

• Merge pull request #9 from AkibaAT/bug/multi-byte-characters
• indexing: panic on multi-byte UTF-8 in bounded chunk truncation

Full Changelog: v0.12.4...v0.12.5

Features

• retrieval: parallel deep search with weighted RRF and query decomposition

Full Changelog: v0.12.3...v0.12.4