Skip to content

1.0.12

Latest
Latest

Choose a tag to compare

View all tags
@EuleMitKeule EuleMitKeule released this 30 Dec 05:01
· 1 commit to master since this release
1.0.12
06bb07e

Release v1.0.12 - Initial Release 🎉

Speaker Recognition brings voice-based speaker identification to Home Assistant, allowing you to identify who is speaking using voice embeddings and the Resemblyzer library.

Features

  • Speaker Recognition Integration: Custom Home Assistant integration that adds speaker recognition capabilities to your smart home
  • RESTful API Service: Standalone API server for voice embedding generation and speaker identification
  • Home Assistant Add-on: Easy-to-install add-on for running the speaker recognition service
  • Speech-to-Text Support: Integration with Home Assistant's STT platform for seamless voice processing
  • PyPI Package: Available as hass-speaker-recognition for easy installation and integration

What's Included

  • Custom Component (speaker_recognition): Home Assistant integration with config flow support
  • Add-on: Dockerized service for local speaker recognition processing
  • Python Package: Installable via pip install hass-speaker-recognition
  • API Client: Python client library for interacting with the speaker recognition service

Installation

Home Assistant Add-on

  1. Add the repository to your Home Assistant add-on store
  2. Install the "Speaker Recognition" add-on
  3. Start the add-on
  4. Add the Speaker Recognition integration via the UI

Python Package

pip install hass-speaker-recognition

Requirements

  • Home Assistant 2023.8 or later
  • Python 3.9+ (for development)
  • Architecture: amd64/x86_64 only (ARM/aarch64 not currently supported due to dependency constraints)

Known Limitations

  • Platform Support: Currently only supports amd64 architecture
  • Python Version: Server features require Python < 3.10 due to Resemblyzer compatibility
  • LLVM Requirement: Requires LLVM 14+ for building dependencies

Configuration

The add-on exposes the following configuration options:

  • host: API server host (default: "0.0.0.0")
  • port: API server port (default: 8099)
  • log_level: Logging level (info, debug, warning, error, critical)
  • access_log: Enable/disable access logging
  • embeddings_dir: Directory for storing speaker embeddings

API Endpoints

  • POST /train - Generate voice embedding from audio
  • POST /recognize - Identify speaker from audio
  • GET /health - Health check endpoint

Full Changelog: https://github.com/eulemitkeule/speaker-recognition/commits/v1.0.0

Assets 4
Loading