LoreAI

Your first document is free, up to 3 pages. No card needed.

Claude OCR: can Claude read images, scans and handwriting?

Yes, Claude reads the text in images and PDFs you upload, up to 20 files per chat[1]. Anthropic says it reads text and visuals only in PDFs of 100 pages or fewer, and text only from 101 pages up, which matters for a long scan with no text layer[1]. For a scan you want back as an editable Word file, LoreAI does OCR on every page.

How to use Claude for OCR

  1. On claude.ai, upload the image like any file, or drag it into the chat window[2]. You can also paste from the clipboard[1].
  2. Ask for the full text of the page, line by line, or as a table.
  3. Copy the answer from the chat.

Claude takes JPEG, PNG, GIF and WebP images, and PDF, DOCX and other documents[1]. For the best reading, Anthropic recommends images of 1000 by 1000 pixels or larger[1].

What limits does Anthropic list?

Files per chat
Up to 20[1]
File size
500 MB per file in a chat, 30 MB per project file[1]
Images
Up to 8000 by 8000 pixels[1], 20 per message and 10 MB each on claude.ai[2]
PDF pages
Up to 1,000 pages[1]
Scanned PDFs
Visual elements read only in PDFs of 100 pages or fewer[1]

That last line matters for OCR. A scanned PDF is made of page images. Anthropic's help page says that for PDFs from 101 to 1,000 pages, Claude processes text only and does not analyze visual elements[1]. A 150-page scan with no text layer gives it nothing to read, so split it into parts of 100 pages or fewer before you upload it.

Can Claude read handwriting?

Claude reads handwriting in a photo as part of the image, the same way it reads print. Anthropic's vision guide lists the cases where it can make mistakes: low-quality, rotated, or very small images under 200 pixels[2]. A sharp, upright photo of each page gives it the best start.

For a whole notebook of lecture notes, LoreAI reads every page in one upload and returns one Word file in the original order, with headings and lists kept. See handwriting to text and the best handwriting OCR.

Claude and LoreAI side by side

Checked on November 9, 2026, against Anthropic's help center and vision docs
What you needClaudeLoreAI
File typesJPEG, PNG, GIF, WebP, plus PDF, DOCX and other documents[1]PDF, JPG, PNG and HEIC
Size500 MB per file in a chat[1]Free 3 MB, Basic 10 MB, Pro 25 MB
LengthPDFs up to 1,000 pages[1]Free 3, Basic 30, Pro 100 pages in one file
Scanned pagesRead as images in PDFs of 100 pages or fewer[1]Every page read with OCR, in order
PhotosUploaded, dragged or pasted into the chat[2]Uploaded directly, HEIC included
What you getAn answer in the chatAn editable Word file (.docx) with headings, lists and tables
TranslationAsk in the chatInto 105 languages in the same step

When to use Claude, and when a document OCR

  • Use Claude to ask about a page: what a chart shows, what a clause means, what a set of notes says in short.
  • Use Claude when the text is a step toward an answer, not the thing you keep.
  • Use LoreAI when you need the document back as a document, every line in order and the tables as tables, ready to edit in Word.
  • Use LoreAI for a long scan in one upload, up to 100 pages on Pro.

For translating with Claude rather than reading, see Claude for translation. The other assistants are compared in ChatGPT OCR and Gemini OCR.

Questions people ask

Can Claude read text from an image?

Yes. Upload the image on claude.ai and ask for the text[2]. Up to 20 files fit in one chat[1].

Can Claude read handwriting?

It reads handwriting as part of the image. Anthropic notes it can make mistakes on low-quality, rotated or very small images[2].

Can Claude OCR a scanned PDF?

For PDFs of 100 pages or fewer it analyzes text and visuals. From 101 pages it processes text only[1], and a scan without a text layer has none.

Is Claude good for converting notes to Word?

Claude answers in the chat, and you paste the text into Word yourself. LoreAI returns the Word file directly, with headings and lists kept.

Sources

  1. Claude Help Center: Uploading files to Claude, accessed November 9, 2026
  2. Claude Docs: Vision, accessed November 9, 2026

By Matt Mountain ยท

Get an editable translation of your document

Upload a scan or a photo. LoreAI returns a Word file in the same layout, ready to check, correct and sign.