What's in. What's next.

Local CLI search with `dubgrep PATTERN [PATH...]`

Optional cache prewarming with `dubgrep index [PATH...]`

Text search across:

Case-insensitive substring matching

Field-restricted search via `--field`

BPM filtering via:

Energy-onset v2 + variable tempo segments

Key detection (Camelot / OpenKey) with confidence scores

FLAC support (full metadata + audio analysis)

`cache status` command with runtime info and library stats

`--max-workers` for controlling indexing parallelism

Plain text output in grep-friendly single-line format

JSON output with `query`, `results`, and `errors`

SQLite cache under `~/.dubgrep/`

Metadata cache write-back during ordinary text search

SHA-256 content identity for cache reuse across renames and moves

Tempo analysis cached by content hash

Tempo model stores segments plus derived `primary_bpm`

Reuse of cached tempo analysis on repeated BPM searches

Reuse of cached metadata on repeated text searches

Supported formats:

Case-insensitive format handling such as `.mp3` / `.MP3`, `.wav` / `.WAV`, `.flac` / `.FLAC`

Recursive scanning via `-r` / `--recursive`

Hidden-file support via `--hidden`

Index summary output with per-file failures sent to stderr

Automated tests covering command behavior, discovery, cache, metadata, decoding, search, and tempo logic