Every photo you own, described and in the right folder.
MediaAnalyzer reads your whole library with a vision model, writes a searchable description next to each file, and sorts it into your folders. From about $1.00 per thousand photos.
curl -fsSL https://mediaanalyzer.pro/install.sh | sh$1 of free credit on sign-up. No card needed.
How it works
Four steps, and you stay in control of the last one.
- 01
Point it at a folder
Run one command on your library, or drop a folder on the web app. It finds every photo and video, however deep.
- 02
Only thumbnails leave
Each file is shrunk to a small preview on your machine. Originals never upload. Metadata and GPS are stripped on arrival.
- 03
AI describes each one
A vision model writes two to four searchable sentences, tags, and picks the one folder it belongs in from your list.
- 04
Your library, sorted
The CLI writes a description file next to each item and moves it into its folder. You review the plan first, and every move can be undone.
A description next to every file
Plain text you can grep, index, or open in anything. It stays yours if you stop using us.
Videos too
Frames are sampled across each clip into a single contact sheet, so a 2-hour video costs the same as a photo.
Your folders, not ours
Edit the folder list to match how you think: kids/soccer, receipts/2026, whatever. The AI picks from it.
Bring your own key
Connect Claude, OpenAI, OpenRouter or your own Ollama/vLLM. Your provider bills the model; we charge a tiny queue fee.
CLI, web, API, DiskPush
A terminal dashboard for big libraries, a web uploader for quick jobs, a REST API, and a DiskPush plugin.
Private by design
Previews are deleted the moment they are described. We keep the text, never the pictures.
A 40,000-photo library for about $40.
Or bring your own model key and pay us about $8 for the queue.