hobokenchicken
0e74a16b40
- Backend: added transcribe_audio (gpt-4o-transcribe), switched audio handler to full blob -> REST -> LLM -> streaming TTS - Frontend: MediaRecorder (webm/opus) full recording sent on stop (one blob per utterance) - Removed dead WhisperStream callbacks and pending_transcript/lock - This unblocks voice per AUDIT item 1 (Option B fallback). Deltas will come in later item. - Also preps for deprecation fix (MediaRecorder is the good path).
78 lines
2.0 KiB
Python
78 lines
2.0 KiB
Python
"""Kira — AI body double backend
|
|
|
|
Realtime WebSocket STT (gpt-realtime-whisper) → gpt-5.4-nano → streaming TTS
|
|
"""
|
|
|
|
import json
|
|
import base64
|
|
import uuid
|
|
import logging
|
|
import time
|
|
import asyncio
|
|
|
|
from fastapi import FastAPI, WebSocket, WebSocketDisconnect
|
|
from fastapi.middleware.cors import CORSMiddleware
|
|
|
|
from config import settings
|
|
from services.memory import kira_memory
|
|
# from services.whisper_stream import WhisperStream # REST fallback active
|
|
|
|
logging.basicConfig(level=logging.INFO)
|
|
logger = logging.getLogger("kira")
|
|
|
|
app = FastAPI(title="Kira Backend")
|
|
|
|
app.add_middleware(
|
|
CORSMiddleware,
|
|
allow_origins=["*"],
|
|
allow_credentials=True,
|
|
allow_methods=["*"],
|
|
allow_headers=["*"],
|
|
)
|
|
|
|
BASE_SYSTEM_PROMPT = (
|
|
"You are Kira, a warm, kind, and encouraging AI body double. "
|
|
"You speak in a friendly, girly-pop tone. You are helping someone with ADHD "
|
|
"stay focused and on task. Keep responses short, supportive, and uplifting. "
|
|
"Check in on them. Remind them to take breaks. Celebrate small wins. "
|
|
"Use occasional emoji but don't overdo it. Never be judgmental."
|
|
)
|
|
|
|
_openai = None
|
|
|
|
|
|
def get_openai():
|
|
global _openai
|
|
if _openai is None:
|
|
from openai import AsyncOpenAI
|
|
_openai = AsyncOpenAI(api_key=settings.openai_api_key)
|
|
return _openai
|
|
|
|
|
|
@app.on_event("startup")
|
|
async def startup():
|
|
if kira_memory.init():
|
|
logger.info("Honcho memory initialized")
|
|
else:
|
|
logger.info("Honcho memory not configured")
|
|
|
|
|
|
@app.get("/api/health")
|
|
async def health():
|
|
mem_status = "active" if kira_memory.enabled else "disabled"
|
|
return {"status": "ok", "name": "kira", "memory": mem_status}
|
|
|
|
|
|
@app.websocket("/api/ws")
|
|
async def conversation_ws(websocket: WebSocket):
|
|
await websocket.accept()
|
|
session_id = str(uuid.uuid4())[:8]
|
|
user_id = "default-user"
|
|
identified = False
|
|
memory_suffix = ""
|
|
logger.info(f"[{session_id}] WebSocket connected")
|
|
|
|
conversation_history: list[dict] = []
|
|
|
|
|