Commit Graph

8 Commits

Author SHA1 Message Date
hobokenchicken c5cc4dd480 fix: replace PCM16 capture with MediaRecorder (Opus/webm)
PCM16 capture via AudioContext was streaming raw audio continuously,
causing massive accumulated buffers that took ~20s to transcribe.
Replaced with MediaRecorder which records compressed Opus/webm and
sends a single blob on release — much smaller, faster to transcribe.

Also removed all unused PCM16/WAV helper functions from both frontend
and backend.
2026-06-04 14:04:44 -04:00
hobokenchicken a370f1ebff fix: use max_completion_tokens for gpt-5.4-nano
gpt-5.4-nano uses max_completion_tokens instead of max_tokens
2026-06-04 13:57:05 -04:00
hobokenchicken a19ac46312 fix: wrap PCM16 in WAV container before STT API call
Frontend captures PCM16 mono 24kHz audio. The transcription API
expects a proper audio container format (wav, webm, etc.), not raw
PCM16 data. Added _pcm16_to_wav() to wrap the raw bytes in a WAV
header before sending to gpt-4o-mini-transcribe.
2026-06-04 13:55:05 -04:00
hobokenchicken f2a5416408 feat: cheapest pipeline — gpt-4o-mini-transcribe + gpt-5.4-nano + TTS
Simple 3-step chat completions pipeline at ~/usr/bin/bash.019/min total.
Streams PCM16 audio from frontend, transcribes on release,
generates response via gpt-5.4-nano, speaks via OpenAI TTS.

Cost breakdown:
  gpt-4o-mini-transcribe: /usr/bin/bash.003/min
  gpt-5.4-nano:          ~/usr/bin/bash.001/min
  OpenAI TTS (nova):     /usr/bin/bash.015/min
  Total:                 ~/usr/bin/bash.019/min (~/usr/bin/bash.57/day at 30min)
2026-06-04 13:51:35 -04:00
hobokenchicken 274d04ea10 feat: hybrid pipeline — gpt-realtime-whisper + gpt-5.4-nano + TTS
Hybrid approach gives streaming STT at ~/usr/bin/bash.017/min + cheap brain
at ~/usr/bin/bash.001/min + TTS at ~/usr/bin/bash.015/min = ~/usr/bin/bash.033/min total.

- gpt-realtime-whisper handles streaming transcription with VAD
- gpt-5.4-nano handles response generation (chat completions)
- OpenAI TTS (nova) for voice output
- Server VAD detects utterance boundaries
- Honcho memory context injected into system prompt
- Removed old full Realtime relay service
2026-06-04 13:48:06 -04:00
hobokenchicken e2332af8d0 feat: OpenAI Realtime API pipeline
Replaced the 3-step sequential pipeline (Whisper STT → DeepSeek LLM
→ OpenAI TTS) with a single OpenAI Realtime API WebSocket using
gpt-4o-mini-realtime-preview.

- ~300-800ms latency vs 1-3s
- Server VAD for automatic turn detection
- Streaming audio chunks during playback
- Interruptions: user can speak over Kira mid-response
- Honcho memory still injected into session instructions
- Frontend captures PCM16 mono 24kHz via AudioContext
- Backend relays client ↔ OpenAI Realtime API
- Supports both voice (PCM16) and text input
2026-06-04 13:32:39 -04:00
hobokenchicken 78ea059f08 feat: user personalization with Honcho-backed preferences
- WelcomeScreen: first-time name entry with cute onboarding
- identify WS message: sets user_id, loads saved prefs from Honcho
- set_preference WS message: saves scene/outfit/accessory to Honcho metadata
- Preferences auto-load on return visits via localStorage + Honcho peer meta
- Kira uses the user's name in greeting and prompts
- Backend: get/set preference methods in KiraMemory service
- Frontend: optimistic preference updates, synced to backend on change
2026-06-04 11:00:58 -04:00
hobokenchicken 97424cb98f init: Kira — AI body double with Honcho memory
Full voice pipeline (Whisper STT -> DeepSeek LLM -> OpenAI TTS),
animated SVG avatar (Live2D-ready), girly-pop UI, lofi music,
timer/notes/pets/wardrobe widgets, 10 background scenes with
particle effects, Honcho cross-session memory.
2026-06-04 10:51:38 -04:00