For the complete documentation index, see llms.txt. This page is also available as Markdown.

Voice Service

Facephi Voice Service is a REST API service to which audio files can be submitted for processing and the results of the speech recognition process retrieved. The service provides one endpoint for enrolling a new voice and another for authenticating a voice.

The product is designed for speaker verification and voice liveness detection. It is based on the use of a voice template, a text string containing the voice’s biometric information. This voice template can be used to authenticate voices in the future.

Optional JWT authentication can be enabled at startup in config.json or through the environment variables FACEPHI_VOICE_REST_AUTH_*. When JWT is enabled, GET /api/v1/version and GET /api/v1/health remain public, while protected endpoints require a valid JWT. The JWT startup configuration is never exposed through GET /api/v1/config and cannot be modified through POST /api/v1/config.

The supported audio formats are:

  • WAV

  • MP3

  • Opus/OGG

  • AAC

  • WMA

  • PCM ulaw and mulaw

  • FLAC

  • ALAC (mov)

  • MP4

  • AIFF

The main distribution contains the following resources:

  • Docker container. The container is available in the FacePhi Docker repository and is based on Ubuntu 24.04. It includes the service and all the dependencies required to run it.

  • Online documentation. The documentation is available at the following URL: https://doc-voice-service.facephi.dev/

Last updated