github LrGenius/LrGeniusAI v3.5.0
Release v3.5.0

4 hours ago

This update brings 2 new features, 1 fix and 10 other changes.

Breaking change

  • The SigLIP2 model has been changed in 3.4.0. You need to download the new AI model, and re-index your catalog (only smart search not AI metadata). Also the AI edit engine has to be retrained manually, since those are bound to the image embeddings generated by the SigLIP2 model.

This however enables LrGeniusAI to improve the smart search as well as introducing a aesthetic scorer model which helps in culling and much more.

New

  • Lightroom flags and cull strength; noise measured on flat regions (@2hkc47w4rd-ux)
  • Browser search with live results and previews

Fixed

  • Link clang_rt on macOS for ort 2.0.0-rc.14

Other changes

  • Aesthetic score and genre for agents; fix find_photos rating/flag/label/file type
  • Index cleanup: take deleted photos out of the search index, automatically
  • Agent chat: a local model works the catalog through the MCP tools, in a browser page
  • Agent chat: keep a conversation's cache between turns (llama.cpp and MLX)
  • Agent chat on cloud models and AI servers, through their own tool calling
  • Enhance MLX model discovery: require chat templates for valid models
  • Agent chat: choose the model on the chat page; get_photo_details with every metadata field
  • Backend access token: every request from the plugin carries one
  • AI Edit: optional aesthetic-score refinement loop
  • Arm64 again and llama.cpp images (Vulkan/CUDA); open --db-path at start; flaky agent-bridge test

Download

If you are on Download
Windows (64-bit) LrGeniusAI-windows-x64-3.5.0.exe
macOS (Apple silicon) LrGeniusAI-macos-arm64-3.5.0.pkg
Your own server or Docker LrGeniusAI-plugin-docker-backend-v3.5.0.zip

Already have LrGeniusAI installed? This update changes parts that the in-Lightroom updater cannot replace on its own, so please install it with the full installer above. Your catalog, settings and everything already analysed stay as they are.

Back up your catalog first

LrGeniusAI writes its results into your Lightroom catalog, and Lightroom counts
each of those writes as an edit: every photo LrGeniusAI processes gets a new
Edit Date, smart collections that filter on Edit Date pick it up, and the
old date cannot be set back. With Automatically write changes into XMP
turned on, Lightroom also writes the new metadata into XMP sidecars and
DNG/JPEG files, so cloud or NAS sync will upload those files again.

Back up your catalog before you run a new version over it, and start with a few
photos. How to back up, and what else to check:
https://github.com/LrGenius/LrGeniusAI/wiki/Getting-Started#before-you-start-back-up-your-catalog

If your computer warns about the download

Both installers are downloaded rarely enough that Windows and macOS may not
recognise them yet. This is about how well known the file is, not about
anything found in it.

  • Windows — if SmartScreen says "Windows protected your PC", click
    More info, then Run anyway.
  • macOS — if the installer is blocked, open System Settings → Privacy &
    Security
    and click Open Anyway. In Terminal, xattr -d com.apple.quarantine <path to the .pkg> does the same thing.

Something not working? Please open an issue at
https://github.com/LrGenius/LrGeniusAI/issues — a description of what you did
and what happened is enough to get started.

Technical changelog

What's Changed

  • ci: bump the actions-minor-and-patch group with 2 updates by @dependabot[bot] in #429
  • MCP: aesthetic score and genre for agents; fix find_photos rating/flag/label/file type by @bmachek in #432
  • Index cleanup: take deleted photos out of the search index, automatically by @bmachek in #435
  • Agent chat: a local model works the catalog through the MCP tools, in a browser page by @bmachek in #433
  • Agent chat: keep a conversation's cache between turns (llama.cpp and MLX) by @bmachek in #434
  • Agent chat on cloud models and AI servers, through their own tool calling by @bmachek in #436
  • Enhance MLX model discovery: require chat templates for valid models by @bmachek in #439
  • Agent chat: choose the model on the chat page; get_photo_details with every metadata field by @bmachek in #440
  • Backend access token: every request from the plugin carries one by @bmachek in #442
  • AI Edit: optional aesthetic-score refinement loop by @bmachek in #441
  • ci: bump the uv group across 1 directory with 2 updates by @dependabot[bot] in #438
  • feat(cull): Lightroom flags and cull strength; noise measured on flat regions by @2hkc47w4rd-ux in #431
  • deps: bump the cargo-minor-and-patch group in /server-rs with 5 updates by @dependabot[bot] in #428
  • fix(lrg-ml): link clang_rt on macOS for ort 2.0.0-rc.14 by @bmachek in #445
  • build(deps): bump the uv group across 1 directory with 2 updates by @dependabot[bot] in #446
  • fix(release): link clang_rt unbundled and run release tests without LTO by @bmachek in #447
  • feat(search-page): browser search with live results and previews by @bmachek in #451
  • Docker: arm64 again and llama.cpp images (Vulkan/CUDA); open --db-path at start; flaky agent-bridge test by @bmachek in #452

New Contributors

Full Changelog: v3.4.0...v3.5.0

Don't miss a new LrGeniusAI release

NewReleases is sending notifications on new releases.