init

Dockerfile

@@ -0,0 +1,20 @@
+FROM python:3.10-slim
+
+RUN apt-get update && apt-get install -y \
+    build-essential \
+    ffmpeg \
+    curl \
+    && rm -rf /var/lib/apt/lists/*
+
+# Install Python dependencies
+WORKDIR /app
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy app and download model
+COPY app ./app
+COPY download-model.sh .
+RUN chmod +x download-model.sh && ./download-model.sh
+
+EXPOSE 4002
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "4002"]

app/main.py

@@ -0,0 +1,19 @@
+from fastapi import FastAPI, File, UploadFile
+import os
+import shutil
+from whispercpp import Whisper
+
+app = FastAPI()
+
+model_path = "./app/model/ggml-base.en.bin"
+whisper = Whisper(model_path)
+
+@app.post("/transcribe")
+async def transcribe_audio(audio: UploadFile = File(...)):
+    temp_file = f"temp_{audio.filename}"
+    with open(temp_file, "wb") as buffer:
+        shutil.copyfileobj(audio.file, buffer)
+
+    text = whisper.transcribe(temp_file)
+    os.remove(temp_file)
+    return {"text": text}

download-model.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+mkdir -p app/model
+cd app/model
+
+# You can choose other sizes too: tiny.en, base.en, etc.
+MODEL_NAME="ggml-base.en.bin"
+MODEL_URL="https://huggingface.co/ggerganov/whisper.cpp/resolve/main/$MODEL_NAME"
+
+if [ ! -f "$MODEL_NAME" ]; then
+  echo "Downloading $MODEL_NAME..."
+  curl -L -o $MODEL_NAME $MODEL_URL
+else
+  echo "$MODEL_NAME already exists."
+fi

files.cli

@@ -0,0 +1,7 @@
+whisper_cpp_api/
+├── app/
+│   ├── main.py
+│   └── model/
+├── Dockerfile
+├── requirements.txt
+└── download-model.sh

requirements.txt

@@ -0,0 +1,4 @@
+fastapi
+uvicorn
+python-multipart
+whispercpp