feat(coding-agent): implemented OpenAI Realtime API voice mode

- Added VoiceSupervisor and VoiceController classes for realtime voice mode using OpenAI Realtime API.
- Added echo suppression, noise floor filtering, and fallback transcript handling for voice input.
- Changed voice mode toggle from Caps Lock to Ctrl+Y with auto-send on silence behavior.
- Changed default TTS model from gpt-4o-mini-tts to tts-1.
- Added kwaipilot/kat-coder-pro model and removed deprecated OpenRouter models.
This commit is contained in:
can1357
2026-01-05 15:11:28 +01:00
parent c34e7cbc2f
commit 8b38705146
13 changed files with 1487 additions and 206 deletions
+4 -2
View File
@@ -42,6 +42,7 @@
"@oh-my-pi/pi-agent-core": "workspace:*",
"@oh-my-pi/pi-ai": "workspace:*",
"@oh-my-pi/pi-tui": "workspace:*",
"@openai/agents": "^0.3.7",
"@sinclair/typebox": "^0.34.46",
"ajv": "^8.17.1",
"chalk": "^5.5.0",
@@ -55,12 +56,13 @@
"nanoid": "^5.1.6",
"ndjson": "^2.0.0",
"node-html-parser": "^6.1.13",
"sharp": "^0.34.2",
"smol-toml": "^1.6.0",
"strip-ansi": "^7.1.2",
"sharp": "^0.34.2",
"winston": "^3.17.0",
"winston-daily-rotate-file": "^5.0.0",
"yaml": "^2.8.2"
"yaml": "^2.8.2",
"zod": "^4.3.5"
},
"devDependencies": {
"@types/diff": "^7.0.2",