Skip to content

LrGenius/LrGeniusTagAI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 

Repository files navigation

LrGeniusTagAI

A Lightroom Classic plugin that uses AI vision models to automatically generate titles, captions, alt text, and keywords for your photos — directly inside Lightroom, on the Library grid or during export.

Features

  • Batch analysis — select any number of photos in the Library grid (or run it as an export action) and let an AI model describe them all in one pass.
  • Multiple AI providers — choose per-install which model does the work:
    • Google Gemini (Flash Lite, Flash, Pro) via the Gemini API
    • OpenAI ChatGPT (Nano, Mini, standard) via the OpenAI API
    • Ollama — run a local vision model, no cloud/API key needed
    • LM Studio — run a local vision model, no cloud/API key needed
  • Configurable output — independently enable/disable generation of title, caption, alt text (for accessibility), and keywords.
  • Keyword hierarchy — optionally sort generated keywords into categories (Activities, Buildings, Location, Objects, People, Moods, Sceneries, Animals, Vehicles, ...) and/or nest them under a top-level keyword per provider (e.g. "Google Gemini").
  • Custom prompts — edit the system prompt/instructions and save multiple named prompt presets to switch between.
  • Multi-language output — generate results in English, German, French, Spanish, or Italian.
  • Context-aware — optionally submit existing GPS coordinates, keywords, or folder names to the AI as extra context, or add free-text photo context per-batch via a dialog before analysis.
  • Review before save — optional validation dialog lets you approve, edit, or reject AI-generated results before they're written to the catalog.
  • Preflight & cost estimate — an optional preflight dialog shows what will be sent before a batch runs, and a running token/cost estimate is shown afterwards (Gemini/ChatGPT pricing tables included; no warranty on accuracy).
  • Debug & performance logging — built-in log file plus optional per-photo CSV performance logging (model, duration, settings) for tuning your setup.
  • Update checks — optional periodic check for new plugin releases, or check manually from the plugin settings.

How it works

  1. The plugin exports a temporary, resized JPEG copy of each selected photo (configurable long-edge size and quality — smaller/lower-quality exports cost less and process faster).
  2. The image (plus any opted-in metadata like GPS/keywords/folder name) is sent to the configured AI provider's vision API along with your prompt.
  3. The AI's structured response is parsed into title/caption/alt text/keywords.
  4. After optional review, the results are written back to the photo's Lightroom metadata, and the temp export file is deleted.

Requirements

  • Adobe Lightroom Classic (SDK 11.0+, i.e. Lightroom Classic 11 / 2022 or later)
  • One of:
    • a Google AI Studio API key (for Gemini models)
    • an OpenAI API key (for ChatGPT models)
    • a local Ollama or LM Studio install running a vision-capable model

Installation

  1. Download LrGeniusTagAI.lrdevplugin from the Releases page (or clone this repository).
  2. In Lightroom Classic, go to File → Plug-in Manager, click Add, and select the LrGeniusTagAI.lrdevplugin folder.
  3. Open the plugin's settings (still in the Plug-in Manager) to configure your AI provider and API key(s) — see Configuration below.

Configuration

All settings live under File → Plug-in Manager → LrGeniusTagAI:

  • AI model — pick the provider/model to use; local Ollama/LM Studio models are auto-discovered if those services are running.
  • API keys — enter your Gemini and/or ChatGPT API key, or set the base URL for a local Ollama/LM Studio server (defaults to localhost:11434 / localhost:1234).
  • Prompts — edit the default system prompt or add/delete named presets.
  • Result language — choose the output language.
  • Generate — toggle which fields to generate (caption, alt text, title, keywords) and whether to review results before saving.
  • Submit existing metadata — opt in to sending GPS, existing keywords, or folder names as extra context to the AI.
  • Keyword hierarchy — enable/disable categorized keywords and edit the category list; optionally nest keywords under a top-level, per-provider keyword.
  • Export settings — long-edge size (px) and JPEG quality used for the temporary image sent to the AI.

Usage

  1. Select one or more photos in the Library grid.
  2. Run Library → Plug-in Extras → Analyze photos with AI (also available as an export action).
  3. Review the preflight dialog (if enabled), wait for the batch to process, and optionally confirm/edit each result in the validation dialog.
  4. Generated title/caption/alt text/keywords are written to the catalog.

Supported AI providers

Provider Models Requires
Gemini 2.5 Flash-Lite/Flash/Pro, 3.1/3.5 Flash-Lite/Flash/Pro Google API key
ChatGPT 5.4 Nano/Mini, 5.4, 5.5 OpenAI API key
Ollama any locally installed vision model Ollama running locally
LM Studio any locally installed vision model LM Studio running locally

Cloud provider usage is billed by the provider according to your own API plan; local Ollama/LM Studio models are free but require sufficient local hardware.

Links