Extract editable text from an image. Online OCR needs no model download; optional on-device OCR supports English and Turkish.
Start with your phone or existing camera. If you need a separate USB camera, this is one option to compare.
You do not need to buy anything to use this free tool.
As an Amazon Associate I earn from qualifying purchases.
A USB webcam with a privacy shutter and video capture up to 1080p at 30 fps.
Check your operating system and USB connection. For document capture, keep the camera steady and use good lighting; this is not a dedicated document scanner.
Links open external sites. Check the exact model and seller in Amazon results; prices and availability can vary by country.
Your image is sent through KitForma to the displayed provider only after your explicit consent.
Extract editable text from a photographed or scanned page. Recognition runs in the cloud after your consent; image preparation stays in your browser.
Checking recognition availability…
Your original appears here
One clear page works best.The preview shows the original. PNG and JPEG are sent unchanged unless you rotate them. WebP and rotated images are converted to PNG; no sharpening or text reconstruction is applied.
Historical and Ottoman-script recognition is not supported yet. Arabic language support does not imply reliable Ottoman transcription. Handwriting and damaged print need careful review.
Your editable text will appear here.
Paragraphs and right-to-left text are preserved when returned by the provider.The link opens this page only. Your files, inputs and private results are not shared.
Choose a related tool for your next task. These links do not transfer your file or inputs automatically; select your file in the next tool.
Choose a clear photo of a printed paragraph.
Online OCR allows limited initial use without sign-in; service availability and separate consent are required. It sends the image to the displayed provider and is subject to the limits shown in the tool. It does not download recognition models to your device. Optional on-device OCR supports printed English/Turkish and downloads language data on first use (15 MiB, 20 megapixels). Recognition can omit or misread text; confidence is not a guarantee. Historical scripts including Ottoman Turkish have not been validated.
Online OCR sends your image through KitForma to the named provider after explicit consent. Limited initial use does not require sign-in. This mode does not download a model to your device. KitForma does not save your image or extracted text in its OCR database. The provider’s data-processing terms apply. Separately selected on-device OCR downloads language models and does not upload the image.
Try a clear image of a short printed sentence you created, without private information. In online mode, check provider availability and the separate consent; in on-device mode, allow the initial language-model download to finish. Compare the result with the source and correct names, numbers and line breaks.
Searchable PDF with OCR: Recognize English and Turkish scans locally and create a PDF with searchable text.
↗Extract PDF text: Save selectable text from PDF pages as a text file.
↗Crop a rectangle using exact pixel coordinates.
↗Count words, characters and reading time.
↗Allow optional analytics cookies? Your files and tool inputs are excluded.
Google Analytics + PostHog + Cloudflare Web Analytics
Tool starts, completions, errors and first download requests are measured. Sent: tool, language and a random temporary attempt ID. Inputs and results are never sent.
With your permission, KitForma also measures public pages, only general section names for private areas such as accounts, and selected action categories on its own server. These analytics are visible only to the site owner. A random tab identifier renews after 30 minutes of inactivity or at most 24 hours; measurement records are kept for up to 31 days. Account IDs, entered content, files, query text and private page addresses are not sent. Withdrawing consent stops these consented measurements.
A separate server counter keeps daily public HTML page access totals even when analytics is declined, without cookies, browser storage or person/session identifiers. This counter does not store IP addresses; network processing for page delivery still occurs. Anonymous counter scope and 31-day retention
With consent, a guide-to-tool click can be linked to that attempt’s result using only public guide/tool codes and a random click code. Fixed campaign labels and recognized referring-source categories can also be used; referring search terms and full addresses are excluded. This context stays in the same tab for at most 30 minutes; filenames, search text and account identity are excluded. Withdrawing consent clears it.
PostHog measures tool usage and technical error categories without personal content, and checks whether optional feedback is enabled. Survey answers are sent only when you submit them. Session replay and automatic click/form tracking are off; account details, files and entered content are excluded.
After consent, Cloudflare Web Analytics measures public page visits and loading times without analytics cookies. Account, support, billing and Etsy areas, URLs containing queries or fragments, and analytics tests are excluded. DNT/GPC signals block collection.
PostHog uses a random identifier stored only for this tab’s session, renewed after 30 minutes of inactivity or at most 24 hours. Withdrawing consent clears it and stops new measurement. An earlier Google Analytics choice does not grant permission for PostHog.
Privacy details