Add Kokoro TTS FastAPI server

Dockerfile

@@ -0,0 +1,21 @@
+FROM python:3.11-slim
+
+WORKDIR /app
+
+# Install system dependencies for soundfile / libsndfile
+RUN apt-get update && apt-get install -y \
+    libsndfile1 \
+    && rm -rf /var/lib/apt/lists/*
+
+# Install Python dependencies
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy application code
+COPY app.py .
+
+# HuggingFace Spaces requires port 7860
+EXPOSE 7860
+
+# Start the FastAPI server
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -1,10 +1,37 @@
 ---
-title: Kokoro Tts Api
-emoji: 🦀
-colorFrom: green
-colorTo: green
+title: Kokoro TTS API
+emoji: 🎙️
+colorFrom: purple
+colorTo: indigo
 sdk: docker
 pinned: false
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Kokoro TTS API
+
+A lightweight REST API for [Kokoro TTS](https://huggingface.co/hexgrad/Kokoro-82M) — used as a serverless backend for the TwelveReader audiobook app, specifically to support iOS Safari where local WASM inference isn't viable.
+
+## API
+
+### `POST /tts`
+Generate speech from text.
+
+**Request body (JSON):**
+```json
+{
+  "text": "Hello world",
+  "voice": "af_bella",
+  "speed": 1.0
+}
+```
+
+**Response:** `audio/wav` binary
+
+### Available voices
+American Female: `af_bella`, `af_heart`, `af_nova`, `af_sky`, `af_sarah`, `af_nicole`, `af_alloy`, `af_aoede`, `af_jessica`, `af_kore`, `af_river`
+
+American Male: `am_adam`, `am_echo`, `am_liam`, `am_michael`, `am_onyx`, `am_puck`, `am_fenrir`, `am_eric`, `am_santa`
+
+British Female: `bf_alice`, `bf_emma`, `bf_isabella`, `bf_lily`
+
+British Male: `bm_daniel`, `bm_fable`, `bm_george`, `bm_lewis`

app.py

@@ -0,0 +1,82 @@
+import io
+import numpy as np
+from fastapi import FastAPI, HTTPException
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import StreamingResponse
+from pydantic import BaseModel
+import soundfile as sf
+from kokoro import KPipeline
+
+app = FastAPI(title="Kokoro TTS API")
+
+# Allow requests from any origin (your Vercel frontend)
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["POST", "GET"],
+    allow_headers=["*"],
+)
+
+# Load the pipeline once at startup — reused for all requests
+# 'a' = American English. Add more pipelines if you want other languages.
+print("Loading Kokoro TTS pipeline...")
+pipeline = KPipeline(lang_code='a')
+print("Kokoro TTS pipeline ready!")
+
+
+class TTSRequest(BaseModel):
+    text: str
+    voice: str = "af_bella"
+    speed: float = 1.0
+
+
+@app.get("/")
+def root():
+    return {"status": "ok", "service": "Kokoro TTS API"}
+
+
+@app.get("/health")
+def health():
+    return {"status": "healthy"}
+
+
+@app.post("/tts")
+async def tts(request: TTSRequest):
+    if not request.text or len(request.text.strip()) == 0:
+        raise HTTPException(status_code=400, detail="Text is required")
+
+    if len(request.text) > 5000:
+        raise HTTPException(status_code=400, detail="Text too long (max 5000 chars per request)")
+
+    try:
+        # Generate audio chunks from Kokoro
+        audio_chunks = []
+        for _gs, _ps, audio in pipeline(
+            request.text,
+            voice=request.voice,
+            speed=request.speed
+        ):
+            audio_chunks.append(audio)
+
+        if not audio_chunks:
+            raise HTTPException(status_code=500, detail="No audio generated")
+
+        # Combine all chunks into a single float32 array
+        combined = np.concatenate(audio_chunks)
+
+        # Encode as WAV and return as a streaming binary response
+        buf = io.BytesIO()
+        sf.write(buf, combined, samplerate=24000, format="WAV", subtype="PCM_16")
+        buf.seek(0)
+
+        return StreamingResponse(
+            buf,
+            media_type="audio/wav",
+            headers={"Content-Disposition": "attachment; filename=audio.wav"}
+        )
+
+    except HTTPException:
+        raise
+    except Exception as e:
+        print(f"TTS Error: {e}")
+        raise HTTPException(status_code=500, detail=str(e))

requirements.txt

@@ -0,0 +1,5 @@
+kokoro>=0.9.2
+soundfile>=0.12.1
+fastapi>=0.110.0
+uvicorn>=0.29.0
+numpy>=1.26.0