This is a simple Streamlit web app that allows you to record speech audio (WAV) based on a list of text prompts, useful for creating speech corpora.

• Reads text prompts from metadata.csv
• Records audio from your browser microphone
• Trims silence using FFmpeg
• Saves audio as WAV files into the wavs/ directory
• Playback support for recorded files

.
├── app.py              # Main Streamlit app
├── metadata.csv        # Prompt list with filename, text, and optional description
├── wavs/               # Output directory for recorded WAV files
├── requirements.txt    # Python dependencies
└── Dockerfile          # Optional Docker build file

Each line should be:

filename.wav|author|lang|Text prompt

Example:

audio1.wav|john|EN|Hello, how are you?
audio2.wav|john|EN|The weather is nice today.

pip install -r requirements.txt
streamlit run app.py

Then open http://localhost:8501 in your browser.

docker build -t corpus-recorder .
docker run -p 8501:8501 -v $(pwd)/wavs:/app/wavs corpus-recorder

• streamlit
• pydub
• ffmpeg (installed system-wide or via Docker)
• streamlit-audiorecorder

MIT License