Imaging labOn-device imaging / Creative tools
Image captions & OCR
Generate a short image description or extract visible text. Review, copy, or download the result without uploading your photo.
Photos stay on your device
Your image
Maximum 16 megapixels. Nothing is uploaded.
Read your image
Describe the scene or extract its text. You can edit the result.
First-use download Up to 620 MB
Models download only when you start an action. Runtime files are extra. Downloads come from Hugging Face and a CDN, and may stay in your browser cache. A local CPU fallback may need another download.
Choose a photo to open the workspace.
Processing stays in your browser.
About this tool Privacy, tips and model credits
Made to stay local.
Your photo is processed on this device. Model downloads need a connection. Your images never leave your browser.
Review, then make it yours.
AI can miss text or invent details. Check names and numbers before using the result.
Good tools. Shared openly.
Powered by Florence-2 vision model. Speed depends on your device. Model credits & licenses.
Tool guide
Describe an image or extract its text
Use one image for two different tasks. Describe image produces a short draft caption for the visible scene. Extract text reads lettering in the image so you can copy it, edit it, or save it as a text file.
Both results are starting points. Review captions before using them as alt text, and verify names, numbers and small print in OCR output. Text hidden by glare, blur or unusual lettering may be missed.
How to use it
- Choose a JPG, PNG or WebP image.
- Select Describe image for a caption or Extract text for OCR.
- Edit the result if needed, then copy it or download a .txt file.
Common questions
Can I use the caption as alt text?
Yes, as a draft. Edit it to describe the image in the context of your page and remove guesses or unnecessary detail.
Does OCR preserve document layout?
No. It returns extracted text, not a formatted document. Check reading order when the image has columns or complex layout.