valentinfrlch/ha-llmvision v1.0.1

v1.0.1: Video Support

on GitHub

Changelog

Features

• Added video support!
  the new video_analyzer service accepts video files as input so you can analyze your security recordings and more. This provides the LLM with even more context which is especially useful for recognizing actions.
• Support for Frigate Recordings
  • Frigate recordings can now be used as input using their event ids.
• #35 Added support for HTTPS requests to self hosted AI providers.
  • Fixed handshake_failed error after setting up local provider with the new HTTPS option

Bug fixes

• Fixed errors not being properly raised