DocoMatic

PDF remediation

Bulk PDF remediation for public entities, verified page by page

School districts, cities, counties and special districts publish thousands of PDFs that were never built to be read with a screen reader.

DocoMatic rebuilds their structure — tags, reading order, tables, forms, alt text — for WCAG 2.1 AA(opens in new tab) and PDF/UA(opens in new tab), then validates the result with veraPDF(opens in new tab) from Level 2 up.

You are not charged for a document that fails verification. Reserved credits are released automatically, and you can re-run the document or choose a higher level.

Scan my whole siteCheck one PDF freeStart free trial — 100 credits

The site scan and the single-file check need no account.

One page, before and after remediationTwo versions of the same page, side by side. On the left, the PDF as published: anonymous blocks of text and images with no structure a screen reader can use. Between them, the veraPDF PDF/UA-1 check. On the right, the remediated file as a tag tree — Document, H1, P, Table with header and data cells, a Figure with alt text, and a list.
As published: no structure
veraPDF PDF/UA-1
DocumentH1PTableTH · TDFigure alt=…L › LI
After: a tag tree

Is this you

Is this you

If three of these are true, you are on the right page.

  • You have a compliance date and no count of how many PDFs you publish.
  • Board packets, agendas, minutes, budgets, permit applications and forms — going back years.
  • Nobody on staff is an accessibility specialist, and nobody is going to be hired as one.
  • You have been quoted per page by a vendor and the number made the project impossible.

If you do not yet know how many documents you publish, start with document monitoring — finding the backlog comes before remediating it. If the problem is the website itself rather than the documents on it, remember that documents are one part of the document accessibility platform.

The problem

The backlog math nobody budgets for

A typical school district or city publishes thousands of PDFs — board packets, agendas, forms, newsletters — and many fail basic accessibility checks. Fixing a backlog that size is a real budget line, and remediation is usually priced per page.

Remediation with verification is built to bring that cost down. Run your own numbers in the remediation cost calculator, or see how the math works for school districts and cities and counties.

  • $0.20–$0.30

    Price per credit on our published monthly plans.

  • 1 · 4 · 30

    Credits a page at Level 1, Level 2 and Level 3.

  • $0.20–$9.00

    Per page, set by what is actually on the page.

Read → rebuild → prove

What happens to every page

Every document runs through nine stages, in three phases: read the page, rebuild its structure, then produce the file and the evidence that it is sound.

Every stage persists its output, so a failure resumes from the last good stage rather than restarting your 400-page board packet from the beginning.

A document on a conveyor passes three archways: a scan grid, a tag-tree scaffold, an orange check seal

Read

Work out what is actually on the page.

Rebuild

Write the structure a screen reader needs.

Prove

Produce the file and the evidence.

  1. Read

    1. Layout analysis

    Every page is decomposed into regions and columns, and candidate reading orders are computed.

  2. Read

    2. OCR where needed

    At Levels 2 and 3, scanned pages get English-language OCR so the text layer matches the visual page. Low-confidence recognition is flagged in the report.

  3. Rebuild

    3. Structure tree

    Headings, paragraphs, lists, tables with header associations, figures, footnotes and table-of-contents entries are built into a proper tag tree.

  4. Rebuild

    4. Artifacts

    Headers, footers, page numbers and decorative elements are marked as artifacts so screen readers skip them.

  5. Rebuild

    5. Alt text and long descriptions

    At Levels 2 and 3, a vision model drafts contextual alt text, with a longer description for charts and complex figures. Very small images are marked decorative. Low-confidence alt text is flagged in the verification report.

  6. Rebuild

    6. Forms

    Form fields get labels, tooltips, tab order and required-field flags so they can be completed with a keyboard and screen reader.

  7. Rebuild

    7. Metadata

    Title, language, viewer preferences and the XMP PDF/UA identifier are set so the document announces itself correctly.

  8. Prove

    8. Tagged PDF written

    The remediated, tagged PDF is written together with a change manifest of everything that was modified.

  9. Prove

    9. Verify

    The analyzer scores the output and veraPDF validates it against PDF/UA-1; at Levels 2 and 3 both must pass. Below threshold, reserved credits are released and you are not charged.

Untagged → tagged

Before and after: what a screen reader sees

An untagged PDF is a picture of a page. The tag tree on the right is what makes headings navigable, tables understandable and forms fillable with assistive technology.

Before — untagged page: anonymous blocks with no structure, order or labels
After remediation, the same page becomes:
After — the tag tree DocoMatic writes
  • Document
  • H1
  • P
  • Table
  • TR › TH scope=col
  • TR › TD
  • Figure alt=…
  • L › LI
  • Form TU=…

Per page, in credits

Three levels, matched to what is on the page

When you upload a document we analyze what it contains and recommend a level. You see the recommendation, its reasons and the credit estimate, and you choose the level before any work runs. Moving a document to a higher level costs the difference in credits. The full table lives on the pricing page.

Credits a page, by levelA bar chart drawn to scale. A Level 1 page uses 1 credit, a Level 2 page 4 credits, and a Level 3 page 30 credits.Level 11 creditLevel 24 creditsLevel 330 credits

  • Level 1

    1 credit per page

    A plain page of text lines and one heading, nothing else on it
    Recommended when the document has
    Text, headings and lists only: no form fields, tables, multi-column pages, images that need alt text or scanned pages
    Typical documents
    Letters, notices, minutes, single-column reports
    Checked against
    Machine checks: tag tree, title, language and no untagged content. Not validated against PDF/UA.

  • Level 2

    4 credits per page

    A two-column page with a table grid and a framed figure, crossed by a scan beam
    Recommended when the document has
    Data tables, multi-column pages, or figures and images that need alt text (a text description for screen readers). Scanned pages also land here, with the OCR add-on.
    Typical documents
    Budgets, agendas, newsletters, staff reports, scanned records
    Checked against
    The Level 1 checks, a DocoMatic Score of 95 or higher, and veraPDF PDF/UA-1 validation with zero failures

  • Level 3

    30 credits per page

    A form page of empty fields, checkboxes and radio buttons, one field with an orange focus ring
    Recommended when the document has
    Fillable form fields. Level 3 adds the forms work: radio button groups, choice lists, required fields, format hints, signature fields and tab order across pages. OCR is included.
    Typical documents
    Permit applications, registration forms, IEP and 504 forms
    Checked against
    The Level 2 checks, plus a reviewer spot-check

The threshold is two separate checks

Our analyzer
The DocoMatic Score is our own measurement of the output, from 0 to 100. Level 2 and Level 3 require 95 or higher. We wrote it and we grade against it, and we would rather say so than present it as independent evidence.
veraPDF
veraPDF(opens in new tab) validates the file against PDF/UA-1 and must return zero failures. It is an independent open-source validator maintained by the PDF Association and the Open Preservation Foundation. We did not write it and we cannot tune it.

The second check is the one your auditor can reproduce. It is also what the no-charge promise is measured against. Level 1 output is structurally checked; PDF/UA-1 validation is not part of its threshold. The authoritative threshold for each level is the table on the guarantee page.

What you can send

What you can send us

What happens to each kind of file, and what OCR costs when a page needs it.

Supported inputs and what happens to each
InputWhat happens
PDF (born-digital)Straight into layout analysis. This is the shortest path through the pipeline.
Scanned PDFEnglish OCR at Level 2 and Level 3, then the normal pipeline. OCR adds 2 credits a page at Level 2 and is included at Level 3. Level 1 does not run OCR.
PNG page imagesAssembled into a document, then OCR, then the normal pipeline.
JPEG page imagesAssembled into a document, then OCR, then the normal pipeline.

Blank pages are detected and not billed. Send documents one at a time, as a batch, from an Amazon S3 bucket, a watch folder or a signed webhook — see integrations.

Honest about the limits

What this doesn't do

Scope, stated plainly, so you can rule us in or out in a minute rather than a month.

  • English OCR only. Scanned pages in other languages are not recognised today. If you publish in Spanish or Chinese, tell us — it affects how we sequence language support.

  • PDFs and page images, not source documents. We do not remediate Word, Excel or PowerPoint files. Remediate the published PDF, or fix the source and re-export — our guide to choosing between Acrobat, checkers and automation covers the trade-offs.

  • Not a certification. No government body certifies documents as accessible, and we do not issue certificates. We remediate to a technical standard and hand you the evidence. Whether that meets your legal obligations is a question for your counsel.

  • Alt text is drafted, not judged. A vision model writes it and flags what it is unsure of. Whether the description means the right thing in your context is a human judgement — that is what human review is for.

  • Not braille, large print or tactile formats. Some entities must provide them, and we do not produce them.

Not legal advice. DocoMatic publishes this as general information for accessibility, records and IT teams. For decisions about your entity's obligations, rely on the primary sources linked here and the DOJ materials at ADA.gov, and consult your attorney.

The report

Verified output, with the report to prove it

Download a sample report (PDF, 62 kB)

A real Level 2 job: score 40 before, 100 after, and 106 veraPDF rules passed with none failed.

A report with a gauge dial and an orange-ringed seal being filed into a shelf of folders

Every remediated file carries a verification report you can download as PDF and JSON: the score before and after, every check that ran, the veraPDF result and a summary of what changed. Keep it for records requests and audits.

Two things are reported, and they are not the same kind of claim. One is our analyzer's score. The other is an independent validator's result.

The DocoMatic Score is ours — we wrote it and we grade against it, and Level 2 and Level 3 must reach 95. veraPDF is not ours: it is maintained by the PDF Association and the Open Preservation Foundation, it validates against PDF/UA-1, and it must return zero failures. Level 1 output is structurally checked and is not PDF/UA-validated.

What it costs

Per-page pricing that fits public budgets

A Level 1 page uses 1 credit, a Level 2 page 4 and a Level 3 page 30. At published plan rates that is $0.20–$9.00 a page, set by what the page contains.

You are not charged for a document that fails verification. Reserved credits are released automatically, and you can re-run it or choose a higher level.

See plans and per-page pricingEstimate your cost

FAQ

Frequently asked questions

Buying for a larger organization?

Book a 20-minute demo(opens in new tab)

How much of PDF remediation is automated?

The pipeline runs without you: layout analysis, tagging, reading order, tables, forms, alt text, metadata and validation. What automation cannot do is judge meaning — whether alt text says the right thing, or whether a reading order makes sense to a person. Where the model is unsure, the item is flagged in the verification report rather than quietly accepted, and you can send the document to human review. We would rather show you the uncertainty than average it away.

Which standards does DocoMatic remediate to?

WCAG 2.1 AA and PDF/UA (ISO 14289). Every file is scored by our analyzer. Level 2 and Level 3 files must score 95 or higher and pass veraPDF PDF/UA-1 validation with zero failures; Level 1 files are structurally checked and are not PDF/UA-validated.

What happens if a document fails verification?

You are not charged for a document that fails verification. Credits are reserved when a job starts and released if the output misses its level's threshold. You can re-run the document, or move it to a higher level and pay the difference in credits.

Can DocoMatic handle scanned PDFs?

Yes, at Levels 2 and 3. Scanned pages get English-language OCR; at Level 2 each scanned page carries an OCR charge of 2 credits on top of the level, and at Level 3 OCR is included. Level 1 does not run OCR. Blank pages are detected and not billed.

When should I add human review?

For documents that must hold up in an audit: forms, legal notices, election materials and high-traffic documents. Human review adds 10 credits a page. Turn it on when you request a fix, or arrange it with our team.

What does a Level 3 page actually look like?

A page with fillable form fields — a permit application, a registration form, an IEP or 504 form. Level 3 adds the forms work: radio button groups, choice lists, required fields, format hints, signature fields and tab order across pages, and OCR is included. It is 30 credits a page, which is why the recommendation matters: you see it, with its reasons, before any work runs. The full table is on the pricing page.

What if a document fails twice?

You are not charged for a document that fails verification, however many attempts it takes. A re-run is free for 30 days after the original job; after that a re-run costs 25 percent of the level's credits. You can also move the document to a higher level and pay the difference.

Can you work from our document management system?

Documents can reach us one at a time, as a batch, from an Amazon S3 bucket, a watch folder or a signed webhook. If your system can write to one of those, it can feed DocoMatic. The integrations page lists what is supported today.

Do you handle documents in languages other than English?

OCR is English-only today. Born-digital PDFs in other languages go through the structural pipeline — tags, reading order, tables and forms do not depend on the language — but we do not currently verify non-English alt text. Tell us what you publish; it affects how we sequence language support.

Start with a count

Bring your backlog to verified accessibility

Start with 100 free trial credits, or check a single PDF right now with no account. If you do not yet know how many documents you publish, scan your site first — that answer is free too.

Start free trial — 100 creditsBook a 20-minute demo(opens in new tab)
Check one PDF freeScan my whole site