v0.4.0
Added
- Built-in digest templates —
meeting-minutes,key-summary(the default digest, structure moved from SKILL.md into the template),analysis-report; the conversational digest prompt now lists the template menu - User templates — drop a markdown file in
~/.config/audio-tldr/templates/: same name overrides a built-in, a new name extends the menu; a format described in conversation can be saved as a reusable template - Digest via subagent — on platforms with subagents the Phase 2 digest runs on a cheaper model by default (Claude Code:
sonnet; Codex:GPT-5.6 Terra); newdigest_modelpreference (a model name pins it,off= digest inline); falls back to inline digesting when dispatch is unavailable - README (en + zh-TW): Digest templates section,
digest_modelrow, update-safety note - 9 new offline template tests (63 total);
transcribe.pyunchanged
Your preferences, custom templates (~/.config/audio-tldr/), and cache live outside the skill folder — updating never touches them.