Image to text converter for photos of signs, pages, labels and posters
You have a photo of a page, a poster, a product label or a whiteboard and you need the words as text you can edit, search or translate. Retyping takes ages. Your todo.is agent reads the image with OCR and AI vision, keeps the structure, and flags any word it could not read with confidence.
The prompt
- Read the text in the attached image(s): [ATTACH THE IMAGES]. The text is in [LANGUAGE]. Give me all of it, keeping [STRUCTURE]. If a word is blurry or cut off, write your best guess followed by a question mark, never invent text. If there are several images, keep them in order and mark each one. Then [EXTRA]. Send the result in the chat and as a .txt file.
What to change
- [ATTACH THE IMAGES]: Attach JPG, PNG, HEIC, WebP or GIF images. Several at once is fine.
- [LANGUAGE]: E.g. "English", "German and English" or "detect it".
- [STRUCTURE]: E.g. "paragraphs and headings", "the line breaks exactly" or "tables as tables".
- [EXTRA]: E.g. "translate it to English", "put the contact details in a list" or "nothing else".
Example result
- Image to text: 3 photos
- Language detected: German · 640 words · each photo marked
- Photo 1: IMG_4021.jpg (a printed recipe card, taken at an angle)
- Straightened and contrast boosted before reading.
- Omas Apfelkuchen
- Zutaten
- • 250 g Mehl
- • 125 g Butter
- • 100 g Zucker
- • 1 Ei
- • 4 säuerliche Äpfel
- Zubereitung
- Mehl, Butter, Zucker und Ei zu einem glatten Teig verkneten und 30 Minuten kalt stellen. Die Äpfel schälen, vierteln und [einritzen?]. Den Teig in eine Springform drücken ...
- Photo 2: IMG_4022.jpg (back of the card, handwritten note)
- "Mit Zimt bestreuen! 180 °C, ca. 45 Min."
- Photo 3: IMG_4023.jpg (a shop sign)
- Bäckerei Krause · Mo–Fr 7–18 Uhr · Sa 7–[13?] Uhr
- Words to check
- • Photo 1, line 12: "einritzen?" is partly hidden by a thumb
- • Photo 3: the closing hour on Saturday is faded, read as 13
- What you can do next
- Reply "translate everything to English", "make it a Word file" or "turn the ingredients into a shopping list".
How to do it with todo.is
- Copy the prompt and set the language and how much layout to keep.
- Attach the photos in todo.is, or just send them to your agent on WhatsApp or Telegram with the prompt.
- Your agent cleans up each image, reads it and marks words it is unsure about.
- Copy the text from the chat or download the .txt. Ask for a translation or a Word file in the same thread.
Tips for a better result
- Take the photo straight on, in good light, with the text filling the frame. It matters more than camera quality.
- For multi-page documents, a PDF scan gives better results than separate photos.
- Say the language, especially for mixed-language text or names. It cuts recognition errors.
- Asking for unclear words in brackets keeps the text honest. You check two words instead of the whole page.
image to text converter: FAQ
- How do I convert an image to text? Attach the image and ask your agent to read it. It runs OCR and AI vision on the picture and returns the text, keeping paragraphs and lists.
- Can it read text from a photo taken on my phone? Yes. Your agent straightens and sharpens the photo first. Very blurry or dark photos may have unclear words, which it marks.
- Which image formats work? JPG, PNG, HEIC from iPhones, WebP, GIF and TIFF. Uploads can be up to 50 MB each.
- Can it read text in other languages? Yes, including accented Latin scripts, Cyrillic, Arabic and Chinese. Say the language for the best result.
JavaScript is required to use the todo.is app.