OpenDocScan
A document scanner in the browser: photograph a page, straighten it, read the text and build a searchable PDF.
Opens the vendor's own site.
Jobs covered
Open a task to see exactly what the vendor documentation supports.
OCR
Browse all ocr toolsbuilding a searchable PDF from photographed/imported page images with OpenDocScan's bundled English OCR model: recognising English text on-device in the browser, gating pages by word confidence and dictionary checks, and writing an invisible Latin-only text layer (unembedded Helvetica/WinAnsi) behind the page image so the exported PDF/library can be searched. The scope records English OCR and searchable-PDF text-layer generation only; it excludes CJK/other-language recognition, handwriting OCR, arbitrary text extraction APIs, existing-PDF editing, redaction, signing, accessibility validation and hosted-service processing, and notes that pages failing the English gate get no text layer
Where your files are processed
Files stayed on your device when we tested
We supplied a marked PNG on 21 October 2026 and no file left the device. Our check confirmed the tool actually ran: The page counted the imported image as a page, so the scanner took the file in.
The scanner, recognition model and PDF build run in the browser tab. The project's README says the Rust scanning module has no HTTP client, and the earlier review measured an image being processed with the network off after the app loaded; scans are stored in IndexedDB, not uploaded.
An optional account is for credits, not scanning. A page-view counter request to the project's own server is separate from document processing, so 'local' describes the document/OCR path rather than every request the page can make.
What it does
OpenDocScan turns a phone photograph of a page into a scan. It outlines the sheet as you frame it, straightens the perspective so an angled photo comes back rectangular, cleans the lighting, adds pages in any order and builds a multi-page PDF with a text layer that the library can search.
Recognition runs in the browser with a bundled English model, and the working copy of a scan lives in the browser's own storage rather than on a server. It is free with no paid tier, no required account, no watermark and no page limit, and the source is dual licensed MIT or Apache-2.0.
The text recognition is English only and its invisible text layer is Latin script only, which the project documents in its own README: CJK pages are rejected rather than mapped, and characters outside WinAnsi are dropped from the searchable layer. The browser version is the complete product; native Android builds are explicitly early and not yet feature-complete.
No tagged release/version is published, so none is asserted.
Best for
Somebody with a phone and a stack of paper who wants PDFs they can search, and who would rather the pages were never sent to a scanner service.
Worth knowing
Recognition is English only, and the project says what that means in practice: a page of Chinese, Japanese or Korean comes back as Latin-looking strings rather than nothing, so every page is checked against word confidences and the model's own dictionary, and a page that fails gets no text layer and no entry in library search. The invisible text layer is Latin-script only and uses unembedded Helvetica with WinAnsi encoding, so words it cannot encode are dropped rather than mangled.
The searchable part of 'scan to searchable PDF' is therefore an English/Latin-script feature. The library lives in one browser profile; clearing site data clears it, and it does not sync to other devices.
No required account, but an optional account exists for credits; the scanner itself needs none. It cannot edit an existing PDF, read handwriting, redact, sign, extract arbitrary text through a programmatic API, validate accessibility or convert existing PDFs.
Native Android builds are early and not feature-complete. The browser version is the complete product.
Details, and where each comes from
Vendor statements, our own reading of their pages, and the things nobody has established yet, each answered separately. Where we have not confirmed something it says so rather than guessing, and a row nobody has established is greyed out.
- Versions covered
- Web
- Website
- opendocscan.com
- Requires an account
- No
- Watermark on the free tier
- No
- Maximum file size
- Not verified
- Batch processing
- Yes
- Data residency
- Not verified
- Open source
- Yes
- Self-hostable
- Yes
- Licence
-
Dual-licensed MIT OR Apache-2.0, at the user's option, per the project README.
Dual-licensed MIT OR Apache-2.0, at the user's option, per the project README. The GitLab project is open-docscan/OpenDocScan; no company owner or legal entity is documented, so owner remains null.
The browser app is free with no paid tier, no watermark and no page limit; an optional account exists for carrying credits across the project's apps and is not required for scanning.
- Redaction
- No redaction feature
- Programmatic access
- None
- Last availability check
- Responded in 79 ms 21 October 2026
Checked automatically by our own runner, not reported by the vendor. 100% of our last 8 checks got a response.
What the redaction verdict rests on
From the vendor's own documentation, read 14 October 2026. Not tested by us.
What that does not establish
Recorded as none from the feature list rather than from a denial. The controls are import, scan, page order, page shape, compression, enhance, black and white, and export, and neither the site nor the README's own known limits section mentions redaction.
A scanner that turns photographs into a PDF has no content to remove, so this is a finding about scope rather than a claim that they handle redaction well.
What the vendor says
Claims are labelled rather than repeated as fact. Where we have checked something ourselves, it says so and shows when. Checked means we read the vendor's own page and it says what we recorded - not that we tested the behaviour, unless the note says we did.
-
“Everything you scan stays in this browser on this device. No sign-up, no upload, no watermark, no page limit.”
Checked by us - the source says this Read 14 October 2026
Confirmed. On the recorded run the file did not leave the device, and the run is confirmed by the page's own counter reaching one page rather than by an absence of requests.
The single request with a body, POST /api/send, carried no file part and no marker, and the site's privacy policy describes it as a page view counted on their own server. On the offline run the scanner took the file in and counted the page with no network at all, which is the stronger evidence because a tool that works with the radio off cannot have uploaded what it processed.
This is one file, on the import path, on one date; the camera capture path was not separately driven.
Check it yourself: the source, the SHA-256, and how we checked
How we checked
Two methods, and the second is the stronger. First, the scanner was driven with a marked PNG while every outbound request was recorded; the recorder pauses each request before it is sent so the body is readable, and it runs three controls against local fixtures before it will report: a page that uploads, a page that does not, and a form post with no file.
Second, the same image was scanned with the browser's network switched off, after the page had been loaded online so its service worker could cache the app. sourceHash is the SHA-256 of the raw page text as fetched on 2026-10-14, held as rv-opendocscan-com-bbcc6e6d.txt.
-
“FREE · OPEN SOURCE · NOTHING IS UPLOADED”
Checked by us - the source says this Read 14 October 2026
Confirmed, and it is the first open source claim in this index that survived being checked. The repository exists at gitlab.com/open-docscan/OpenDocScan, the README states dual licensed under either MIT or Apache-2.0, and LICENSE-MIT and LICENSE-APACHE both return 200 with the expected first lines.
Free is stated in the header, the pricing section and two FAQs, including that recognition is free, and no pricing tier exists.
Check it yourself: the source, the SHA-256, and how we checked
How we checked
The repository the site links was fetched, the README's licence section was read, and both licence files it names were requested directly. sourceHash is the SHA-256 of the raw page text as fetched on 2026-10-14, held as rv-opendocscan-com-bbcc6e6d.txt.
Each claim is quoted from the vendor's own pages. This section describes what we have and have not checked - it is not a judgement about anyone.
Sponsorship is distribution, not an endorsement or a promise of search rankings, AI citations, traffic, or sales. This listing is a free entry, compiled from published sources.
Related reading
- Which PDF tools keep your file on your machine - what we count as processing on your own device.
- Open source PDF tools, and what their licences require - what open source does and does not tell you.