Image to text
Upload a photo, scan or screenshot and get the text back in reading order. It reads Arabic, Persian, Urdu, Hindi, Chinese, Cyrillic and Latin script, among others. Your first 5 pages each day need no account.
Keep going with a free account
You've used today's free pages. Sign in with GitHub or Hugging Face and get $5 of free credit: about 2,500 pages at preview prices: up to 50 pages on your first day, then up to 500 a day. You also get the playground, where you can compare models, and an API key.
No card needed. The free pages here reset at 00:00 UTC.
Continue in the playground
You've used the free pages here for today. You're signed in, so the playground gives you up to 500 pages a day (50 on your account's first day), with model comparison and your API keys.
Images are read by an open preview model on Cloudflare Workers AI. We don't store them or use them for training. See the privacy notice.
The text will appear here.
Questions
Is it free?
Yes. You can read 5 pages a day here without an account. If you sign in, you get $5 of free credit, which is about 2,500 pages at preview prices (up to 50 on your first day, then up to 500 a day). Nothing is billed during the preview.
Which languages and scripts does it read?
Most widely used scripts, including Latin, Arabic, Persian, Urdu, Hebrew, Cyrillic, Greek, Devanagari, Chinese, Japanese and Korean. Results depend on the image: a sharp, straight photo works best. Rune-1, our own model, is being trained for the harder cases.
Do you keep my images?
No. The image is sent to the model, the text comes back to you, and nothing is saved. We only count how many pages you've used today.
I'm signed in. Does this page use my credit?
No. This page always uses the 5 free pages a day, even when you are signed in. Your credit is used in the playground and through the API: up to 500 pages a day, or 50 on the first day of a new account.
Can I use it from my own code?
Yes. Sign in to the console, create an API key and send images to POST /v1/ocr. See the API docs.
Can it read PDFs?
Not yet. For now, take a screenshot of the page or export it as an image. PDF support comes with Rune-1.