PDFix
On-premises PDF accessibility desktop app and SDK: PDF/UA and WCAG validation, auto-tagging and batch remediation.
Jobs Accessibility Edit
Where your files are processed
No server involved
A desktop application and an embeddable SDK, documented as fully on-premises with no document data sent to PDFix servers or any cloud service. One caveat that belongs with the verdict: two optional AI actions in the vendor's own marketplace do send page images to Amazon Textract and content to OpenAI, so the no-upload position is true of the core product and false for those specific actions. That is a property of the library itself: what an application built with it does with your file depends on that application, not on this.
Architectural, and stated repeatedly by the vendor: Desktop and SDK run on Windows, macOS and Linux, and the SDK is documented as fully on-premises with no document data transmitted to PDFix servers or any cloud service. This is the product's main selling point to US federal customers. The exception is the AI actions marketplace, where the Amazon Textract action sends page images to AWS and the OpenAI action sends content to OpenAI; the vendor documents the Textract case explicitly and calls the local IBM Docling action the alternative if fully offline processing is required.
What it does
PDFix is a desktop application and a developer SDK for PDF accessibility work, sold by a Slovak company with a US arm. The desktop tool validates a PDF against PDF/UA, WCAG 2.1 and 2.2 and Section 508 using the veraPDF engine, then helps repair it: auto-tagging, table and list structure, reading order, alt text, font embedding and batch runs over whole folders. The SDK exposes the same engine as a library for Python, C#, Java, C++ and Node.js so remediation can be built into a document pipeline, and both are documented as processing entirely on the customer's own machines and network.
Best for: Compliance teams, accessibility specialists and engineering teams with sensitive or classified PDFs who must remediate at volume without documents leaving their own infrastructure.
Worth knowing: The free download is not a clean free ride: the vendor's FAQ says saved documents carry a watermark until a licence is bought, so a free tier exists but its output is not production-ready. It also does not remove the need for human judgement, and the vendor says so itself: it does not guarantee 100% document accessibility and the responsibility for compliance rests with the user, so a passing validation run is not by itself proof of compliance. Two of the AI actions in its own marketplace send page images to Amazon Textract or OpenAI, which contradicts a blanket reading of its no-upload position. Redaction is an SDK Professional-tier feature and is not available in the PDFix Desktop application — confirmed by its absence from the Desktop product page, pricing, changelog and all 45 Desktop user-guide pages, and by the SDK pricing table ticking it only in the Professional and Enterprise columns. On licensing, free use is not the same as open source: the Desktop Lite checking tier and the SDK Lite reading and rendering tier are permanently free, but a matching core-library licence for the binary was not substantiated in this review, and an examples repository does not establish the licence of what you ship.
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
- Windows · macOS · Linux · API
- Website
- pdfix.net
- Requires an account
- Not applicable
- Free tier
- Yes
- Watermark on the free tier
- Yes
- Maximum file size
- Not verified
- Batch processing
- Yes
- Data residency
- Self-hosted
- Open source
- Not verified
- Self-hostable
- Yes
- Licence
- Proprietary commercial EULA (PDFix s.r.o.); Desktop Pro and Enterprise licensed per user per year, SDK tiers per instance by quote; SDK Lite licensed under Apache 2.0
- Redaction
- Removes the content
- Programmatic access
- Embeddable library · Command line
- Last availability check
- Responded in 806 ms 16 October 2026
Checked automatically by our own runner, not reported by the vendor. 90% of our last 30 checks got a response.
What the redaction verdict rests on
From the vendor's own documentation, read 22 September 2026. Not tested by us.
“Redaction permanently removes sensitive content from the selected area. … Once the text or graphics are removed, they cannot be retrieved.”
What that does not establish: The vendor’s own feature page, not a test we ran. The scope is narrower than the verdict implies: "from the selected area". The page separately claims automatic removal of all private information and personal data connected to the redacted area, which is a broader claim and is not established by that sentence.
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.
-
“PDFix Desktop and PDFix SDK are fully on-premises - your documents never leave your computer or network. There is no cloud upload or SaaS processing.”
Checked by us — the source says this Read 21 September 2026
Confirmed verbatim. The claim is the first two sentences of the answer to the question "Does PDFix upload documents to the cloud?", which reads: "No. PDFix Desktop and PDFix SDK are fully on-premises - your documents never leave your computer or network. There is no cloud upload or SaaS processing." The rest of the same answer is worth recording because it is the kind of substantiation this index rarely gets: the vendor names USCIS and the Judicial Council of California as users relying on it "for sensitive and classified documents under strict data residency requirements (GDPR, FISMA, HIPAA)", and contrasts itself with "most competitors", which it describes as "Windows-only or web-based SaaS". Named customers and named frameworks are checkable claims, and neither has been checked here. Recorded as the vendor's description of a desktop application and an SDK; we have not run either and no control was run.
Standing: Re-read and matched byte for byte
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- d0a1b4f6a84eabec7884aa310c948b548af93e124f5c078377d23771d9e95b28
How we checked: The FAQ on https://pdfix.net/pricing/ is an accordion that headless Chrome did not render into text — the rendered text is 4868 bytes and contains no match — so the answer was read from the raw HTML response, saved as 246427 bytes and hashed. The sentence appears three times in that response: twice in the markup of the FAQ item and once inside the page's JSON-LD structured data. sourceHash is the SHA-256 of the raw response.
-
“The free PDFix download includes all features - there is no feature restriction between free and paid versions... The only difference is a watermark on documents saved with the free version.”
Checked by us — the source says this Read 21 September 2026
Confirmed verbatim, including the em-dash: "The free PDFix download includes all features — there is no feature restriction between free and paid versions." The answer then lists what that means in practice — "full WCAG 2.1, WCAG 2.2, PDF/UA (ISO 14289) validation, AI auto-tagging, manual remediation, batch processing — same for SDK" — and closes with the exception the claim also carries, the watermark on documents saved with the free version. This is the second PDFix claim read from collapsed markup in this campaign, after the pricing-page FAQ verified earlier, and the pattern is the same: the vendor's most useful sentences about licensing are the ones a reader has to click to see. Recorded as the vendor's description; no download was made and no control was run.
Standing: Re-read and matched once per-response tokens were masked
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- 79620945abe7a61f0eafef07d4fab7a560781e45b9e5e13b032828dcf13057b1
How we checked: The FAQ answer is not in the rendered page — https://pdfix.net/faq/ rendered 10082 bytes of visible text containing no match — because the answers sit in collapsed cards. The claim was read from the raw HTML response, 591595 bytes saved and hashed, where the answer appears in a card headed "What does free PDFix version include?". sourceHash is the SHA-256 of that raw response.
-
“PDFix SDK Lite is permanently free - no registration, no time limit.”
Checked by us — the source says this Read 21 September 2026
Confirmed verbatim: under the question "Is PDFix SDK free to use?" the answer opens "PDFix SDK Lite is permanently free – no registration, no time limit." The same answer scopes it and names what is included, both of which the claim leaves out: the free tier covers "PDF rendering, reading PDF objects, content extraction, and metadata reading", and above it sit commercial tiers — Basic, Professional and Enterprise — priced by instance count with quote-only pricing. So "permanently free" is accurate for Lite and says nothing about the tiers a commercial user would need, which is the same shape as the PDFix desktop claim verified in this batch: free, complete, watermarked.
Standing: Re-read and matched byte for byte
Check it yourself: the source, the SHA-256, and how we checked
- From
- the raw HTTP response
- SHA-256
- a108eef301c11a93ab43ac10f63bd27b004daf4c3508167444d1dd0b70bfb939
How we checked: The recorded source names an FAQ on the SDK pricing page; the answer is not in the rendered text (https://pdfix.net/pricing/sdk/ rendered 5909 bytes with no match) because it sits in a collapsed accordion, so the claim was read from the raw HTML response, 255293 bytes saved and hashed. The sentence appears twice in that response, once in the page's markup and once inside embedded JSON. sourceHash is the SHA-256 of the raw response.
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
- Checking a PDF for accessibility — what makes a PDF accessible, and how it is checked.
- Which PDF tools really redact — which tools remove the content rather than cover it.
- Which PDF tools keep your file on your machine — what we count as processing on your own device.