Interview Audio Model / Voice Analyzer

A lightweight voice analysis service that extracts features, transcribes audio, scores responses, and returns insight metrics for interview-style audio recordings.

Project layout

• app/ — FastAPI application entry and API routes (app/main.py).
• app/api/routes/analyze.py — primary analysis endpoint (app/api/routes/analyze.py).
• services/ — domain services (transcription, scoring, sentiment, features).
• schemas/ — request/response Pydantic schemas.

Features

• Upload audio files and receive transcription, sentiment, scoring, and insights.
• Modular services so you can swap models or feature extractors.
• Docker-ready for easy deployment.

Requirements

• Python 3.10+ (recommended)
• See requirements.txt for full dependency list.

Quickstart (local)

1. Create and activate a virtual environment:

python -m venv .venv
.\.venv\Scripts\Activate.ps1

2. Install dependencies:

pip install -r requirements.txt

3. Run the app (development):

uvicorn app.main:app --reload --port 10000

The API will be available at http://localhost:10000. Automatic API docs (Swagger UI) live at http://localhost:10000/docs.

API

POST /api/analyze

• Description: Analyze an uploaded audio file and return transcription, sentiment, scores, and insights.
• Endpoint implementation: app/api/routes/analyze.py
• Request: multipart/form-data with file (audio file, e.g., WAV/MP3) and optional JSON fields specified in schemas/request_schema.py.
• Response: JSON described by schemas/response_schema.py.

HTTP endpoints exposed by the service

• POST /api/v1/analyze - unified speech intelligence dashboard.
• POST /api/v1/communication/analyze - communication-focused analysis.
• POST /api/v1/emotion/timeline - emotion timeline analysis.
• WS /ws/emotion-stream - live emotion stream for PCM audio chunks.

The React frontend expects the service to be reachable at http://localhost:10000 by default. Override that with VITE_AUDIO_MODEL_URL in the frontend environment if needed.

Example curl

curl -X POST "http://localhost:10000/api/v1/analyze" \
	-F "file=@/path/to/answer.wav" \
	-H "accept: application/json"

Docker

• Build image locally:

docker build -t voice-analyzer:local .

• Run with docker-compose:

docker-compose up --build

Development notes

• Core configuration is in core/config.py.
• Services live under services/ (e.g., services/transcription_service.py, services/scoring_service.py).
• Add or replace ML models in the services directory; keep interfaces consistent.

Contributing

• Open issues or PRs. Follow typical Python project conventions.

License

• Add your license of choice here.

Questions or changes

• Tell me if you want a longer README, badges, or CI instructions.