Document accessibility remediation & NIMAS conversion

Everybody needs an editor.

Machines included. We take the document work that has to be right and can't be done by hand at volume — and hand every judgment call back to a person.

What we do

Assess and remediate documents. Word and PDF files evaluated against applicable WCAG 2.1 AA and Section 508 criteria — including headings, alt text, reading order, contrast, and tables — with remediation and a scoped report that supports ADA Title II obligations and states the limits of testing.
Produce NIMAS filesets. Curriculum source turned into the validated package that braille, large print, audio, and digital text are made from, with print pagination preserved.
Triage a backlog. Point us at a folder or repository. Get back what has no issues detected within tested coverage, what may be suitable for automation, and what needs a specialist — with the evidence and planning estimates behind each route.
Copyedit at scale. Suggestions arrive with the reason attached, bound to the exact run of text. Judgment-based edits remain pending until a person reviews them.
Build it into your shop. For on-premises deployments, the systems can run on hardware, model accounts, and network rules you control. Specs and a test corpus define the agreed boundary.

Start with a quote See how it works

Deployment and data handling are scoped to the engagement

On-premises deployments can keep processing on hardware and networks you control.
Hosted work uses only the systems and data paths authorized for the engagement.
Source files, derived evidence, logs, and backups are retained only as authorized and for an agreed period.
We do not use client content to train models. Any approved third-party processing must use no-training terms.

What we build

Accessibility remediation · NIMAS · conversion

The five above, in detail — four systems and the shop they run in. Everything here exists so that a person who couldn't read the book, can.

PincerPilot workflow

Pincer

DOCX accessibility remediation — WCAG 2.1 AA

Perceivable · operable · understandable

Reads a .docx and checks for known issues such as untagged headings, images with no alt text, tables with no identified header row, and contrast below the configured threshold. It proposes supported fixes with evidence and produces a scoped report that states what was and was not evaluated.

In: .docx, .dotx
Out: .docx + report (HTML, PDF)
Aligned to: Applicable WCAG 2.1 AA · Section 508 criteria

TrawlOperational workflow

Trawl

PDF accessibility audit & backlog triage

Point it at a folder. It sorts each file into no issues detected within tested coverage, an automation candidate, or specialist review — with the evidence and planning assumptions behind each route. Results describe the checks performed; they are not a certificate of compliance.

In: Folder of .pdf
Out: Triage manifest (CSV, JSONL)
Scale: Corpus-dependent · measured during intake

SashPilot workflow

Sash

AI-assisted copyediting — every change approved by a person

The suggestion pass leaves the source untouched. Each proposed edit comes back with a reason and a locator tied to the text it reviewed. Judgment-based changes remain pending until a person decides; stale or ambiguous locators are routed back for review rather than applied blindly.

In: .docx, .xml (DITA), Markdown
Out: Tracked changes + rationale
Rule: Preserve by default

MateProduction service

Mate

NIMAS filesets — braille, large print, audio, digital text

a11y · one source, every reader

A student who reads by touch, by ear, or at 24 point needs the same book as everyone else, on the same day, on the same page. Mate takes curriculum source — InDesign, Word, legacy XML — and assembles the NIMAS fileset those formats are made from: real structure, images with alt text worth reading, MathML where there's math, and print pagination preserved so a class can work together. It validates before it ships and hands you the report, so a rejection is something you find here rather than downstream.

In: IDML, .docx, legacy XML
Out: NIMAS fileset + validation report
Feeds: Braille · large print · audio · digital text

The engine

Our orchestration layer routes each agreed step to the tool suited to it: a model for bounded judgment, a local model for suitable bulk work, and deterministic code for operations that must be exact.

Tooling is selected within the deployment, data-handling, and review boundaries agreed for the engagement. We document the workflow so the result does not depend on a mystery process.

Specific tools and model versions can be documented for each engagement. The evidence package records the processing steps, checks, and decisions included in the agreed scope.

Judgmentalt text, plain-language rewrites, edge cases → a frontier model

Bulktriage, classification, extraction at volume → approved local or hosted processing

Exacttagging, structure, math that must be right → deterministic code

Recordagreed processing steps and review decisions → scoped audit trail

How it behaves

Audit · propose · approve · apply

First

Audit

The system reads your files and identifies issues within the stated test coverage. It changes nothing yet.

Then

Propose

Supported judgment-based fixes come back as suggestions, with a reason and a locator for the content they affect.

Then

Approve

You keep, skip, or edit each one. Skipping is a normal answer, not a failure.

Last

Apply

Approved changes and pre-authorized deterministic rules are written, re-checked, and logged. The original is never overwritten.

Judgment-based changes require human review. Pre-authorized automated rules are documented and validated.

These systems don't replace your editors. They hand your editors a shorter, better-evidenced list — so the judgment stays with the people who were always doing the judging.

Work with us

WCAG 2.1 AA · Section 508 · support for ADA Title II obligations

Start with a quote

Send us a representative slice of the backlog. You get findings, severity, and an honest estimate of the work — including the parts a machine can't do. No fixes, no obligation, no call required. Most engagements start here, and some end here.

Have us build it

We quote a system off your audit, not off a rate card. The number reflects the risk we take off your desk. If the audit says the pile is easier than you feared, the quote says so too.

Run it yourself

The systems on your own hardware, your own model accounts, your own network rules. Includes the specs, the test corpus, and the known limitations within the agreed scope. Licensed per production team.

Tell us what's on your desk. We'll tell you what it takes.

info@finelinesediting.com

We report tested coverage and known limitations. We do not sell a certificate of compliance, a legal opinion, or interactive form remediation.