A LITTLE MODEL. A NEW PERSPECTIVE.
Every photo
has a story.
Choose a moment from your gallery.
Let a little AI put it into words.
Your photos stay on this device.
LET’S GET READY
A little setup. Then your photos.
- Check browser compatibility
- Download and prepare SmolVLM2
- Choose a photo and describe it
The model starts downloading automatically: roughly 350–650 MB. Wi-Fi recommended. Keep this page open. Saved model files are reused on later visits when available.
Checking this browser…
A fresh pair of eyes.
Landscapes, everyday objects, little details.
What will you discover?
JPEG, PNG, WebP, or any image your browser can open.
Put it into words.
SmolVLM2 · 256M parameters
First use downloads roughly 350–650 MB. Wi-Fi recommended. Model files are cached when browser storage allows.
THE DESCRIPTION
AI can miss details or make mistakes. Check the description against your photo.
Made for your phone. Designed for your privacy.
Photos are resized locally and never uploaded. Model files download from Hugging Face. No account, analytics, or API key. Descriptions are in English.
Requires WebGPU: a compatible Android browser or Safari on iOS 26+. Older devices may run out of memory. Keep this app open while it works.
For offline use, finish a description online first so all model files are cached. Browser storage can be cleared or evicted; offline access is not guaranteed. On iPhone, use Share → Add to Home Screen to install.