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.
- 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:
- 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.
- 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).
- 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.
- The AI's structured response is parsed into title/caption/alt text/keywords.
- After optional review, the results are written back to the photo's Lightroom metadata, and the temp export file is deleted.
- Adobe Lightroom Classic (SDK 11.0+, i.e. Lightroom Classic 11 / 2022 or later)
- One of:
- Download
LrGeniusTagAI.lrdevpluginfrom the Releases page (or clone this repository). - In Lightroom Classic, go to File → Plug-in Manager, click Add, and
select the
LrGeniusTagAI.lrdevpluginfolder. - Open the plugin's settings (still in the Plug-in Manager) to configure your AI provider and API key(s) — see Configuration below.
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.
- Select one or more photos in the Library grid.
- Run Library → Plug-in Extras → Analyze photos with AI (also available as an export action).
- Review the preflight dialog (if enabled), wait for the batch to process, and optionally confirm/edit each result in the validation dialog.
- Generated title/caption/alt text/keywords are written to the catalog.
| 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.
- Website & docs: https://lrgenius.com
- Configuration docs: https://lrgenius.com/configuration
- Ollama setup guide: https://lrgenius.com/help/docs/help-ollama-setup/