add a simple python fastapi server to serve whisper models for speech-to-text and piper for text-to-speech

audio_server/.gitignore

@@ -0,0 +1,3 @@
+temporary_audio/
+whisper_models/
+piper_models/

audio_server/.idea/.gitignore

@@ -0,0 +1,5 @@
+# Default ignored files
+/shelf/
+/workspace.xml
+# Editor-based HTTP Client requests
+/httpRequests/

audio_server/.idea/.name

@@ -0,0 +1 @@
+audio_server

audio_server/.idea/audio_server.iml

@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="PYTHON_MODULE" version="4">
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$">
+      <sourceFolder url="file://$MODULE_DIR$/src" isTestSource="false" />
+      <excludeFolder url="file://$MODULE_DIR$/venv" />
+    </content>
+    <orderEntry type="jdk" jdkName="Python 3.14 (audio_server)" jdkType="Python SDK" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>

audio_server/.idea/inspectionProfiles/Project_Default.xml

@@ -0,0 +1,23 @@
+<component name="InspectionProjectProfileManager">
+  <profile version="1.0">
+    <option name="myName" value="Project Default" />
+    <inspection_tool class="InconsistentLineSeparators" enabled="true" level="WARNING" enabled_by_default="true" />
+    <inspection_tool class="IncorrectFormatting" enabled="true" level="WEAK WARNING" enabled_by_default="true" />
+    <inspection_tool class="PyPep8NamingInspection" enabled="true" level="WEAK WARNING" enabled_by_default="true">
+      <option name="ignoredErrors">
+        <list>
+          <option value="N803" />
+          <option value="N806" />
+        </list>
+      </option>
+      <option name="ignoredBaseClasses">
+        <list>
+          <option value="unittest.TestCase" />
+          <option value="unittest.case.TestCase" />
+          <option value="todo_extract.Task" />
+        </list>
+      </option>
+    </inspection_tool>
+    <inspection_tool class="TodoComment" enabled="true" level="WARNING" enabled_by_default="true" />
+  </profile>
+</component>

audio_server/.idea/inspectionProfiles/profiles_settings.xml

@@ -0,0 +1,6 @@
+<component name="InspectionProjectProfileManager">
+  <settings>
+    <option name="USE_PROJECT_PROFILE" value="false" />
+    <version value="1.0" />
+  </settings>
+</component>

audio_server/.idea/misc.xml

@@ -0,0 +1,7 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="Black">
+    <option name="sdkName" value="Python 3.14 (transcription_server)" />
+  </component>
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.14 (audio_server)" project-jdk-type="Python SDK" />
+</project>

audio_server/.idea/modules.xml

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/audio_server.iml" filepath="$PROJECT_DIR$/.idea/audio_server.iml" />
+    </modules>
+  </component>
+</project>

audio_server/.idea/vcs.xml

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="VcsDirectoryMappings">
+    <mapping directory="$PROJECT_DIR$/.." vcs="Git" />
+  </component>
+</project>

audio_server/Containerfile

@@ -0,0 +1,19 @@
+FROM python:3.14.4-slim AS builder
+
+WORKDIR /app
+COPY pyproject.toml requirements.txt ./
+COPY src src
+
+RUN pip wheel --no-cache-dir --no-deps --wheel-dir wheels .
+
+FROM python:3.14.4-slim AS runner
+
+COPY --from=builder /app/wheels /wheels
+RUN pip install --no-cache /wheels/* && rm -rf /wheels
+RUN apt update
+RUN apt -y install ffmpeg
+
+WORKDIR /app
+
+EXPOSE 8000
+CMD ["uvicorn", "audio_server:app", "--host", "0.0.0.0", "--port", "8000"]

audio_server/pyproject.toml

@@ -0,0 +1,20 @@
+[build-system]
+requires = ["setuptools>=82.0.1", "wheel"]
+
+[project]
+name = "audio_server"
+description = "a fastapi server serving whisper and piper"
+version = "1.0.0"
+authors = [
+    { name = 'Milan', email = 'milan.boemer@gmail.com' }
+]
+
+requires-python = '>=3.12'
+
+dynamic = ["dependencies"]
+
+[tool.setuptools.dynamic]
+dependencies = { file = ["requirements.txt"] }
+
+[tool.setuptools.packages.find]
+where = ['src']

audio_server/requirements.txt

@@ -0,0 +1,8 @@
+openai-whisper==20250625
+fastapi~=0.135.3
+psutil~=7.2.2
+uvicorn~=0.44.0
+starlette~=1.0.0
+python-multipart~=0.0.24
+aiofiles~=25.1.0
+piper-tts~=1.4.2

audio_server/src/audio_server/__init__.py

@@ -0,0 +1,76 @@
+import logging
+import os
+import uuid
+from wave import Wave_write
+from pathlib import Path
+
+import aiofiles
+from piper import PiperVoice, SynthesisConfig
+from piper.download_voices import download_voice
+import psutil
+from fastapi import FastAPI, UploadFile
+import whisper
+from pydantic import BaseModel
+from fastapi.responses import FileResponse, JSONResponse
+
+app = FastAPI()
+
+whisper_model_name = os.getenv("WHISPER_MODEL", default="small")
+whisper_path = Path("whisper_models")
+whisper_path.mkdir(parents=True, exist_ok=True)
+whisper_model = whisper.load_model(whisper_model_name, download_root=whisper_path.as_posix())
+
+piper_model_name = os.getenv("PIPER_MODEL", default="de_DE-karlsson-low")
+piper_path = Path("piper_models")
+piper_path.mkdir(parents=True, exist_ok=True)
+download_voice(piper_model_name, piper_path)
+
+temporary_audio_path = Path("temporary_audio")
+temporary_audio_path.mkdir(parents=True, exist_ok=True)
+
+logger = logging.getLogger(__name__)
+
+@app.get("/")
+async def index():
+    return {
+        "memory_usage": psutil.Process(os.getpid()).memory_info().rss,
+        "piper_model": piper_model_name,
+        "whisper_model": whisper_model_name,
+    }
+
+@app.get("/transcribe")
+async def transcribe(audio_file: UploadFile):
+    file_extensions = {
+        "audio/wave": "wav",
+        "audio/mpeg": "mp3",
+    }
+
+    if audio_file.content_type not in file_extensions.keys():
+        logger.error("%s is not supported", audio_file.content_type)
+        return JSONResponse({"detail": "wrong file type"}, status_code=400)
+
+    file_path = temporary_audio_path / f"{uuid.uuid4()}.{file_extensions[audio_file.content_type]}"
+
+    async with aiofiles.open(file_path, "wb") as f:
+        content = await audio_file.read()
+        await f.write(content)
+
+    # noinspection PyArgumentList
+    transcription = whisper_model.transcribe(file_path.as_posix())
+
+    return JSONResponse(transcription)
+
+class VoiceRequest(BaseModel):
+    text: str
+    config: SynthesisConfig | None
+
+@app.get("/tts")
+async def tts(voice_request: VoiceRequest):
+    voice = PiperVoice.load(piper_path / f"{piper_model_name}.onnx")
+
+    audio_file_path = temporary_audio_path / f"{uuid.uuid4()}.wav"
+
+    with Wave_write(audio_file_path.as_posix()) as writer:
+        voice.synthesize_wav(voice_request.text, writer, syn_config=voice_request.config)
+
+    return FileResponse(audio_file_path, media_type="audio/wav")