AI keywording for Lightroom Classic has gone from "barely exists" to a real category in the space of about two years. Adobe never added native AI tagging to Classic (their focus stayed on Develop and performance), which opened the door for third-party plugins. Today there are several legitimate options, and the right one depends entirely on your workflow and priorities.

This is a practical, honest comparison. We make MetaMagic, so we have a horse in this race — but the goal here is to help you choose well, not to pretend the alternatives don't have strengths. If a different tool fits your needs better, you should use it.

What to look for in an AI keyword plugin

Before comparing specific tools, here's what actually matters when evaluating an AI metadata plugin:

  • AI model quality — does it recognize specific species, landmarks and objects, or just generic labels?
  • IPTC support — does it write proper IPTC fields, or just dump keywords?
  • Batch processing — can it handle hundreds or thousands of photos reliably?
  • Setup complexity — does it need your own API key, local GPU, or technical config?
  • Pricing model — free, one-time, subscription, or pay-as-you-go credits?
  • Stock agency support — does it export in formats agencies accept?

The comparison

Feature MetaMagic Self-hosted / open-source plugins Cloud Vision plugins
AI modelOpenAI visionGemini / ChatGPT / local (your choice)Google Cloud Vision
Needs your own API keyNoUsually yesUsually yes
Needs local GPUNoOptional (for local AI)No
Writes IPTC fieldsYes — full IPTCVariesKeywords mainly
IPTC location from GPSYes — automaticRarelyRarely
Title + caption + keywordsAll threeVariesKeywords mainly
Editorial caption modeYesNoNo
Web app (no install)YesNoNo
Stock agency CSV export12 agenciesNoNo
Pricing modelPay-as-you-go creditsFree / one-timeFree / pay-per-use
Setup difficultyMinimalTechnicalModerate

Comparison reflects general categories of tools available in 2026. Specific features of individual plugins change frequently — always check the current version of any tool before deciding.

Who each type is best for

Open-source / self-hosted plugins

Great if you're technical, want full control, and don't mind getting your own API key (or running a local AI model on a capable GPU). The big advantage is cost — if you bring your own OpenAI or Gemini key, you pay the raw API price with no markup. The trade-off is setup complexity and the lack of niceties like agency exports, editorial mode, or a support line when something breaks. If you're comfortable in a terminal and reading GitHub issues, this route can be very economical.

Cloud Vision plugins

Tools built on Google Cloud Vision have been around the longest. They're solid for basic keyword tagging but tend to produce more generic labels — "bird", "building", "car" — rather than the specific identifications that drive stock sales. They usually focus on keywords alone, without titles, captions or location handling.

MetaMagic

Built for photographers who want results without setup. No API key, no GPU, no config — register and start. It uses OpenAI's vision model for specific, accurate identifications, writes full IPTC (title, caption, keywords, location), offers editorial mode, and exports CSV for 12 stock agencies. It's also the only option with both a Lightroom plugin and a browser web app sharing one account. The trade-off versus bring-your-own-key tools is that credits include the AI cost plus the service, so per-image cost is higher than raw API pricing — you're paying for the convenience, the polish and the support.

The honest bottom line

If you're highly technical and cost is your top priority, a bring-your-own-key open-source plugin will be cheaper per image. If you want the best AI quality, zero setup, full IPTC output, editorial support, agency exports, a web app, and a developer who answers emails personally — that's what MetaMagic is built for.

The best way to decide is to try it on your own photos. MetaMagic includes 20 free generations on signup — enough to compare the keyword quality against whatever you're using now, on your own images, before spending anything.

Related reading