Skip to main content
FastOCR

Smallpdf OCR tool review: is it good for scanning, and what to use instead

Short answer

Is Smallpdf good for OCR scanning? For a small PDF OCR job, yes: it makes a scanned PDF searchable with no sign-up. But the free plan has an unpublished daily download limit; our second file was refused with a 6.5-hour wait. On our one scored page, Arabic, its text came back as Latin letters. FastOCR (we make it) returns text, a searchable PDF, Word and Markdown, free for images.

By the FastOCR team · Updated 4 November 2026 · Accuracy test run 31 October 2026

We make FastOCR. Accuracy figures come from our own test of each tool we scored, ours included, on the same scans; other facts link to each tool's own pages. How we tested

Try it here

Drop your file here

PDF, PNG, JPG, WebP, BMP, HEIC

PDF up to 1 GB · images up to 20 MB

Tips for a clear photo
  • Lay the page flat.
  • Use even light, with no shadows.
  • Turn off the flash to avoid glare.
  • Fill the frame with the page.

Verdict

Pick FastOCR if

  • You want to read and copy the text on screen first, then download a searchable PDF, Word or Markdown from the same upload.
  • Your scan is Arabic or another right-to-left script. FastOCR got 2.6% of our Arabic page wrong; Smallpdf got 99.1% (lower is better).
  • You have several images today: with a free account, image OCR has no daily cap.
  • The file must not leave your computer: the free Mac and Windows apps (beta) work offline.

Pick Smallpdf if

  • You already merge, compress and convert files in Smallpdf and want OCR in that same account.
  • You plan to edit the words inside the PDF afterwards; the Pro plan lists Edit Text next to OCR.
  • You only need one or two PDFs made searchable today, in a language you have checked it on.

FastOCR and Smallpdf side by side

Sign-up for free OCRFastOCRImages: none. PDFs: a free account.SmallpdfNot needed. The OCR page says "no signups".Source
Free limit, as advertisedFastOCRA small daily image cap until you sign in; a free account removes it. See pricing.SmallpdfA "daily download limit" on the free plan, with no number given.Source
A second free file the same dayFastOCRImages: works (visitors have a small daily cap; signed in, none). PDFs: works with a free account, within the free monthly pages.SmallpdfRefused with "You’re out of free downloads" and a wait of about 6.5 hours (31 October 2026).Our test
Files it acceptsFastOCRPhotos, screenshots, scanned images and PDFs.SmallpdfPDF, Word, PowerPoint, Excel, JPG, BMP, GIF, TIFF and more.Source
What you get backFastOCRRaw text on screen, a searchable PDF, Word (.docx) and Markdown.SmallpdfA searchable PDF. Word is a second step through its separate PDF to Word converter.Source
LanguagesFastOCR100+ languages, detected automatically, including mixed pages.Smallpdf"Many different languages"; the OCR page does not list them.Source
Arabic page, share of characters wrong (lower is better)FastOCR2.6%Smallpdf99.1%: the text layer was Latin-letter garbage.Our test
Watermark on free outputFastOCRNone on text, Word, Markdown or the searchable PDF.SmallpdfNot statedSource
What happens to uploadsFastOCRProcessed on our servers; see the privacy policy.SmallpdfDeleted after one hour unless you save them to a Smallpdf account.Source
Paid planFastOCRMonthly plans or one-time page packs.SmallpdfSmallpdf Pro, with a 7-day free trial. The price is not stated on the pages we checked.Source

Why FastOCR

  • Scanned PDFs

    Scanned PDF in, searchable PDF out; text copies in the right order, including right-to-left.

    See Scanned PDFs

  • Photos and screenshots

    Every image upload gets raw text and a single-page searchable PDF, plus Word and Markdown.

    See Photos and screenshots

  • Languages

    100+ languages, detected automatically, including mixed pages.

    Share of characters wrong on our evidence scans (lower is better): Arabic 2.6%, Urdu 9.1%, Bengali 7.4%, Tamil 2.2%, Chinese 1.0%.

    See Languages

  • Free use

    Images: free, no sign-up, no watermark. PDFs: free account.

    See Free use

  • Offline desktop apps

    Free offline Mac and Windows apps (public beta): searchable PDF, text, Word and Markdown, nothing leaves your computer.

    See Offline desktop apps

Proof

Smallpdf allowed one free download, so it has one score: 99.1% of the Arabic page wrong, against FastOCR's 2.6% (lower is better). PDF24, scored on the six pages shown, was higher than FastOCR on every one.

Three lines of a printed Arabic book page, a scan from Wikimedia Commons, starting with the words في معية هذا المعلم.(opens full size)

FastOCR

2.6% of characters wrong
في معية هذا المعلم علم ان الامبراطور غورديانوس فتح ابواب هيكل
يانوس في رومة ليعلن الحرب على ساسان . فصمم الفيلسوف الطالب على
الالتحاق بهذه الحملة العسكرية ليسمع عن فلسفة الفرس والهنود. والتحق

Smallpdf

99.1% of characters wrong
u
'H as 'He 'in 'wtf 'woo
\     'a.:;ffu1;{ art; Fu he u.:a2sv5;    aw up   »'ir/d0s'0!.1r,'d Pam 'NT   "r 'Zaqrg
The opening lines of an Arabic book page and the first lines of the text layer in the PDF Smallpdf returned for it, unedited apart from leading spaces trimmed. Each percentage is the share of characters that tool got wrong on the whole page, so lower is better. This is one page: Smallpdf then said it was out of free downloads, so we did not score its other seven. FastOCR output is from 25 October 2026; Smallpdf ran on 31 October 2026. Scan: Wikimedia Commons (public domain). Text: Wikisource.

% of characters wrong by script, lower is better

  • FastOCR
  • PDF24 OCR
  • Smallpdf (Arabic page only)
  • OCR.space Engine 3
  • imagetotext.info
Character error rate by script, lower is better
ScriptCharacter error rate by tool
Chinese
FastOCR 1.0%
PDF24 3.2%
Smallpdf no score
OCR.space E3 no score
Tamil
FastOCR 2.2%
PDF24 9.2%
Smallpdf no score
OCR.space E3 no score
Arabic
FastOCR 2.6%
PDF24 35.5%
Smallpdf 99.1%
OCR.space E3 2.9%
imagetotext.info 2.9%
English
FastOCR 3.3%
PDF24 10.1%
Smallpdf no score
OCR.space E3 no score
Bengali
FastOCR 7.4%
PDF24 18.6%
Smallpdf no score
OCR.space E3 11.2%
imagetotext.info 7.9%
Urdu
FastOCR 9.1%
PDF24 69.0%
Smallpdf no score
OCR.space E3 19.6%
imagetotext.info 9.2%
Character error rate on one public-domain book page per script. Lower is better. Smallpdf was scored on the Arabic page only: it ran out of free downloads after that page. Gaps under about 3 points are within the noise of the reference text. "Not supported" means the tool refused the language; "no score" means it returned nothing, which says nothing about accuracy. How we tested.

Try it on your own page

Is Smallpdf good for OCR scanning?

Smallpdf is good for OCR scanning when you have one scanned PDF and want it back searchable, without making an account. Its OCR tool takes PDF, Word, PowerPoint, Excel and image files, says "no signups", and returns a searchable PDF you download or share. If you need Word, the page sends you on to its separate PDF to Word converter. Uploads are deleted after one hour unless you save them to a Smallpdf account.

Three things are worth knowing before you rely on it. The page says it reads "many different languages" but does not list them, and the tool showed us no language menu. The free plan has a daily download limit with no number on it, and we hit it on our second file. And the one page we could score, an Arabic book scan, came back as Latin letters. For English pages we have no Smallpdf score, so test one of your own.

Smallpdf free OCR: the daily download limit

Smallpdf free OCR stops after a small number of downloads a day, and the number is not published. The OCR page says its tools are "free to use, though some limits apply", and the pricing page marks PDF OCR on the free plan with a "daily download limit". On 31 October 2026, with no sign-in, our first download worked. The second was refused with "You’re out of free downloads", an offer of Smallpdf Pro with a 7-day free trial, and a countdown of 06:34:12.

For one file a day that limit may never matter. For a folder of scans it decides your afternoon. FastOCR also caps images for visitors without an account, at a small daily number; a free account lifts it, so image OCR then has no daily cap. Scanned PDFs use a free monthly page allowance shown on the pricing page.

Smallpdf OCR, second download

The dialog said

“You’re out of free downloads”

“Get Smallpdf Pro to unlock full access and download your file or wait 06:34:12.”

It appeared on our second download, with no sign-in and the free tool, on 31 October 2026.

FastOCR in any browser

  1. 1Open fastocr.org
  2. 2Drop in a photo
  3. 3Copy the text or download it

No account for 3 images a day. Sign in free for unlimited images.

Smallpdf's OCR tool after one free download, 31 October 2026, beside the FastOCR steps for the same job.

Smallpdf vs PDF24: accuracy in OCR

PDF24 has the fuller accuracy record of the two, because Smallpdf stopped us after one page. We ran the same public-domain book pages through PDF24; on the six scripts shown (Arabic, Urdu, Bengali, Tamil, Chinese, English) it got 3.2% to 69.0% of characters wrong. FastOCR got 1.0% to 9.1% on those pages. Smallpdf got 99.1% of the Arabic page wrong, against 2.6% for FastOCR, and its other pages could not be downloaded. Every figure is the share of characters wrong, so lower is better. See also the full benchmark.

One page is a narrow sample. It shows that Smallpdf's automatic language handling missed an Arabic book scan on that day, not how it reads English or Chinese. Try your own page in both tools before a long job.

What FastOCR gives you instead

FastOCR shows the recognised text next to your page, so you can check and copy it before you download anything. Each image comes back as raw text, a single-page searchable PDF, a Word file and Markdown, with no sign-up and no watermark. Scanned PDFs come back searchable, and the text copies in the right order, including right-to-left Arabic, Urdu and Hebrew. It detects 100+ languages by itself, mixed pages included, so there is no language to pick. Once an image is read, you can translate its text or tidy it with AI Polish. The quickest test is the free OCR online uploader on the home page.

Smallpdf is a PDF toolbox with OCR inside. FastOCR is OCR first, with these free extras:

Photo of a page

Photo of an old printed English book page, slightly grey and uneven, starting with the words 'ing, and Haggart gained an opportunity'.(opens full size)

Text FastOCR returned

3.3% of characters wrong
ing, and Haggart gained an opportunity
to escape; he left Perth for Glasgow, and
made up his mind to go to Ireland, and
started in the Steam Boat; here he wa
recognised by the Provost of Kircaldy
who went ashore at Lamlash; Haggar
The first lines of a page FastOCR read on 25 October 2026, shown as it returned them; words cut at the page edge stay cut. The percentage is the character error rate: the share of characters wrong, missing or extra, so lower is better. Scan: Wikimedia Commons (public domain). Text: Wikisource.

How we tested

  • Pages: Wikimedia Commons book scans in the public domain or CC0, one per script. This page shows six of them: Arabic, Urdu, Bengali, Tamil, Chinese, English. The reference is the Wikisource transcription.
  • Score: character error rate, the share of characters wrong, missing or extra after removing spaces, punctuation and Arabic diacritics, so lower is better. The reference differs from our own proofread key by 1% to 6%, so gaps under about 3 points are noise.
  • Smallpdf: the free tool at smallpdf.com/pdf-ocr on 31 October 2026, no sign-in, optional cookies declined. It offered no language menu. We wrapped the Arabic scan in a one-page PDF, the way scanned PDFs usually arrive, and read the returned text layer with pdftotext. The Urdu download was refused as "out of free downloads", so the other pages have no score.
  • PDF24: its free OCR PDF tool on 31 October 2026, no sign-in, language set by hand for each page.
  • FastOCR: output from its production API on 25 October 2026. It got a smaller 500-pixel-wide image for the Arabic, Urdu and English pages; the other tools got 1,280 pixels.
  • Smallpdf's limits, outputs and file handling come from its own pages, checked on 4 November 2026. Scans: Wikimedia Commons. Text: Wikisource.

Choose Smallpdf if

  • You already pay for Smallpdf Pro or use its other PDF tools every week, and want OCR in the same place.
  • You need to edit text inside a PDF, which its Pro plan lists next to OCR.
  • You want uploads removed after one hour by default, unless you choose to save them.
  • You make the odd PDF searchable, in a language you have checked it on, and one file a day is enough.

FAQ

Is Smallpdf good for OCR scanning?

For a single scanned PDF it is a reasonable choice: no sign-up, and a searchable PDF back. In our test its one scored page, an Arabic book scan, was 99.1% wrong (lower is better), and the second free download was refused. Test a page of your own before a long job.

Is Smallpdf OCR free?

Yes, with a cap. The OCR page says the tools are free to use though some limits apply, and the pricing page lists a daily download limit without a number. On 31 October 2026 our second free download was refused, with a wait of about 6.5 hours.

Does Smallpdf OCR need an account?

No. Its OCR page says no signups are needed. An account matters if you want to keep files past one hour or try Smallpdf Pro, which has a 7-day free trial. FastOCR also reads images without an account; PDFs need a free one.

Which languages does Smallpdf OCR read?

Its OCR page says it recognises text in many different languages but does not list them, and the tool showed us no language menu. Run one page in your language before a long file. FastOCR detects 100+ languages automatically, including pages that mix scripts.

Smallpdf vs PDF24: which is more accurate for OCR?

We can only score PDF24 fully: 3.2% to 69.0% of characters wrong across the six pages shown. Smallpdf got 99.1% of the Arabic page wrong, then its free downloads ran out. FastOCR got 1.0% to 9.1% on the same six. Lower is better.

Can Smallpdf turn a scanned PDF into Word?

Yes, in two steps. The OCR tool returns a searchable PDF, and its page points you to the separate PDF to Word converter for a .docx file. FastOCR offers raw text, Word and Markdown downloads from the same upload as the searchable PDF.

What is a free alternative to Smallpdf OCR?

FastOCR. Images need no sign-up and carry no watermark, and each comes back as text, a searchable PDF, Word and Markdown. Scanned PDFs need a free account and come back searchable, with right-to-left text that copies in the right order.

Method

Smallpdf facts come from its own OCR and pricing pages, checked on 4 November 2026; a fact it does not state is shown as Not stated, never as No. Accuracy is the share of characters wrong against the Wikisource text, lower is better, on public-domain scans run on 31 October 2026. Smallpdf was scored on one page only. Details are in how we tested.

Change log

  • 18 July 2026: First published.
  • 31 October 2026: Added our Arabic test of Smallpdf OCR and its free download limit.
  • 4 November 2026: Smallpdf facts re-checked on its own pages; moved to the shared comparison format.