WAVStream is a WebSocket server designed to work with FreeSwitch's mod_audio_stream module. It enables real-time audio call recording, saving as WAV files, and automatic uploading to S3-compatible storage.
- Receive both text and audio data via WebSocket from the FreeSwitch
mod_audio_streammodule - Create WAV files in real-time with filenames specified in the initial FreeSwitch connection command
- Automatic WAV header updating
- Automatic upload to S3-compatible storage upon recording completion
- Support for changing the recording file mid-call
- Concurrent handling of multiple recordings
-
Clone the repository:
git clone https://github.com/ynigun/wavstream.git cd wavstream -
Install dependencies:
go mod tidy -
Create a
.envfile in the project directory and set the following variables:S3_ENDPOINT=your-s3-endpoint S3_ACCESS_KEY_ID=your-access-key S3_SECRET_ACCESS_KEY=your-secret-key S3_BUCKET_NAME=your-bucket-name S3_USE_SSL=true S3_PREFIX=audio/ SERVER_PORT=8080
Start the server:
go run main.go
The server will start running on port 8080 by default.
To start a recording with FreeSwitch, use the following command in the FreeSwitch CLI:
uuid_audio_stream <UUID> start ws://<SERVER_IP>:8080/ws mixed 8k rec_name_uuid.wav
Replace <UUID> with the UUID of the call, <SERVER_IP> with the IP address of the WAVStream server, and rec_name_uuid.wav with your desired filename for the recording.
To stop the recording and trigger the S3 upload:
uuid_audio_stream <UUID> stop
To change the recording file mid-call:
uuid_audio_stream <UUID> send_text rec_name_uuid_B.wav
This will close and upload the current recording to S3, then start a new recording with the specified filename.
This project is distributed under the GNU General Public License v3.0. See the LICENSE file for more details.