PDFFree trial

OCR — image or PDF to text

Extract text from a scanned PDF, screenshot, or photo using AI vision.

Last updated 19 May 2026

Checking your session…

How to use OCR — image or PDF to text

About under a minute, start to finish.

1
Open the OCR tool
Visit /tools/pdf/ocr-pdf.
2
Upload your image
JPG, PNG, WebP, or GIF up to 20 MB. For PDFs, convert pages to JPG first with the PDF-to-JPG tool.
3
(Optional) hint a language
Auto-detect works for most cases; specify when your image has mixed languages or unusual scripts.
4
Click Extract text
Server-side via GPT-4o-mini vision. Most images return text in 5-15 seconds.
5
Copy or download
Plain text output preserving line breaks and table structure. Copy to clipboard or download as .txt.

Frequently asked about OCR — image or PDF to text

How accurate is the OCR?

On clean printed text — typed documents, screenshots of articles — it's near-perfect (>99%). On handwriting it's good but variable (~85-95%). Tables and multi-column layouts are preserved as Markdown.

Why GPT-4o vision instead of Tesseract?

GPT-4o handles handwriting, multi-column layouts, and tables much better than open-source OCR. Trade-off: variable cost per image (~$0.0015) — bundled into your Apexkit monthly medium-meter budget.

Can I OCR a whole PDF at once?

Not in one shot — Whisper-API/vision has a per-call image limit. Convert pages to JPG via the PDF-to-JPG tool first, then OCR each page.

All included in your Apexkit plan. No extra cost.

See all

Chat with PDF

Hero

Ask any PDF questions — get cited answers in seconds.

PDF Summarizer

Hero

Get a 1-paragraph or bullet-point summary of any PDF.

AI Resume Tailor

Hero