v0.3.3

Changes

• Speaker attendance filter (--speakers): Limit which enrolled speakers are matched during identification
  • any2md podcast.mp3 --diarize --identify --speakers "Alice,Bob" — constrain search space
  • Unknown/guest speakers reliably detected when not in the expected set
• Speaker groups: Reusable named speaker sets
  • any2md speaker group create "Hosts" --members "Alice,Bob"
  • --speakers "@Hosts" expands group to member list
  • Full CRUD: create, list, show, delete, add-member, remove-member
• JSON attendance field: --json output includes attendance.expected, attendance.identified, attendance.unknown

v0.3.2

Changes

• Fix speaker deps: wespeaker → wespeakerruntime (PyPI installable, ONNX-based)
  • Removed torch dependency for speaker identification
  • Updated speaker.py to use wespeakerruntime.Speaker(lang='en') API
  • Full catalog pipeline verified: enroll, match, unknown rejection

v0.3.1

CLI standards compliance: shell completions, docs command, llms.txt, --silent flag.

v0.3.0 — Speaker Identification

Speaker Identification via WeSpeaker ResNet293

New feature: named speaker identification across recordings using persistent voice profiles.

New: any2md speaker subcommand

any2md speaker add "Joe" --audio joe-sample.wav
any2md speaker list
any2md speaker remove "Joe"
any2md speaker merge "Speaker A" "Speaker B"
any2md speaker stats "Joe"
any2md speaker gallery "Joe"

New: --identify flag

any2md meeting.m4a --diarize --identify
# Output: **Joe** [00:24] instead of SPEAKER_0

How it works

• WeSpeaker ResNet293 extracts 256-d speaker embeddings (PyTorch MPS on Apple Silicon)
• Gallery model: stores multiple embeddings per speaker to handle voice variation across mics/conditions
• sqlite-vec for fast KNN nearest-neighbor search
• Adaptive thresholds: high-confidence auto-match (≥0.85), medium-confidence with score (0.70-0.85)
• Auto-enrolls new embeddings for matched speakers (gallery grows over time)
• Prompts for unknown speakers (or --auto-enroll / --no-enroll)

Speaker catalog

Persistent at ~/.config/any2md/speakers.db:

• Gallery maintenance: rolling window of 20 embeddings per speaker
• Per-speaker distance statistics for drift detection
• Merge support for duplicate profiles
• Audit trail for profile merges

Install

brew upgrade any2md
# Install speaker identification deps
uv pip install any2md[speaker]

Stats

• 201 new tests (119 speaker + 42 CLI + 40 yt)
• 7 tickets implemented (ANY2-11 through ANY2-17)

v0.2.3

Remove redundant mp3 intermediate conversion

• Local files (m4a/mp4/wav/etc): Skip mp3 step, convert directly to 16kHz WAV (~16s faster on long files)
• YouTube: Skip mp3 postprocessor in yt-dlp, keep native format (opus/webm) for direct WAV conversion — also avoids lossy mp3 re-encoding
• Output filenames: Remove _whisper suffix that was leaking from intermediate files into output markdown filenames

v0.2.2

Fix diarization OOM on long audio files

Breaking change: None — drop-in fix.

What changed

• diarize() now uses model.generate_stream() (chunked streaming) instead of model.generate() (full-file, O(n²) memory)
• Added _merge_diarization_segments() for cross-chunk speaker segment merging
• 9 new unit tests + 2 integration tests with 82-min public domain podcast

Before/After

• Before: 48-min file → [metal::malloc] Attempting to allocate 83404284800 bytes (OOM)
• After: 48-min file → 12.6s diarization (289 chunks, 712 segments, 4 speakers)

Other fixes

• Fixed test_cli.py hardcoded version string
• Updated CLAUDE.md with current line counts

v0.2.1 — repo subcommand

New: repo subcommand

Pack an entire git repository into a single markdown file via repomix.

# Repomix JSON straight to stdout (for agents)
any2md repo ./my-project --json

# Markdown with frontmatter
any2md repo ./my-project -o ~/docs

# Auto-detects directories with .git
any2md ./my-project

# Tree-sitter structure only (compressed)
any2md repo . --compress

Requires npm install -g repomix as a system dependency.

Stats

• 19 subcommands (18 converters + repo)
• 20 new tests for repo subcommand

v0.2.0 — CLI Standards Upgrade

CLI Standards Upgrade

New Features

• --json / -j output mode for all 16 converters — structured JSON to stdout for agent consumption
• --fields flag for JSON field selection (dot-notation: frontmatter.rows,content)
• --version / -V flag
• --quiet / -q flag to suppress log output
• NO_COLOR environment variable support
• Shell completions for all subcommands
• deps subcommand — shows installed/missing optional dependencies
• Structured JSON error output to stderr in --json mode

Improvements

• All log output now goes to stderr; stdout reserved for data
• Standardized exit codes: 0=success, 1=user error, 2=usage error
• Improved error messages for missing optional deps (suggest uv pip install commands)
• Help text footer with bug report URL and homepage
• Added mammoth to doc optional deps (fixes DOCX conversion)

New Files

• llms.txt — agent-readable capability index
• Makefile — standard test/lint/install targets
• WORKLOG.md — development history

Stats

• 9,500+ lines of source across 18 modules
• 766 tests

v0.1.0

Initial release. 16 converters, 740 tests, local MLX inference on Apple Silicon.