Releases: S1M0N38/pi-voice
Releases · S1M0N38/pi-voice
Release list
v3.0.0
3.0.0 (2026-07-31)
⚠ BREAKING CHANGES
- migrate to @earendil-works namespace and pi 0.83 ModelRuntime
Bug Fixes
- cli: run the TS CLI via a jiti launcher, spawn server with absolute jiti-register (84faaf2)
- extension: play audio via execFile with player fallback; temp WAVs in os.tmpdir() (e7e1ebd)
- migrate to @earendil-works namespace and pi 0.83 ModelRuntime (926b6d4)
- register jiti via module resolution for global install (99f3e63)
- restore saved config on /voice reset (35ebc0b)
- server: serialize model ops, self-heal corrupt cache, fix dtype file mapping (93140d6)
- server: strip markdown from text before synthesis (7e1201e)
v2.0.1
v2.0.0
2.0.0 (2026-05-06)
⚠ BREAKING CHANGES
- Config moved from ~/.pi/voice.json to ~/.pi/voice/config.json. The summaryModel field is removed; use per-event model in events config instead. EventConfig now accepts { prompt, model? } or { text } (mutually exclusive).
Features
- overhaul event config and migrate state to ~/.pi/voice/ (3242af6)
v1.2.0
v1.1.0
v1.0.1
v1.0.0
1.0.0 (2026-05-05)
⚠ BREAKING CHANGES
- rename template placeholders to pi-voice
Features
- add Kokoro TTS HTTP server (68302fa)
- add pi extension with tts tool and /voice command (bd17b81)
- add prepare script for default config on install (d2c7415)
- custom TUI, auto-TTS events, and async speak (69c6c05)
Bug Fixes
- voice: use pi-voice CLI syntax in server hints (93cbdad)
Code Refactoring
- rename template placeholders to pi-voice (8398cf8)