Datalab's new pipeline turns scanned PDFs into tagged accessible PDFs in minutes, replacing $5–25/page manual work
VikParuchuri · x · 2026-09-09
Datalab released a new WCAG accessible PDF processing pipeline: input any PDF — including a scan that is pure pixels — and get back a tagged PDF readable by assistive tech, with a real text layer under the scan and a full structure tree (headings, levels, table ownership).
- Why it matters: WCAG is a set of success criteria, not a file format; PDFs are just drawing instructions, and scans are bitmaps, so machines can't know what's a heading or which column a number belongs to.
- Economics: manual tagging costs $5–$25 per page at published vendor rates, taking a day or two for simple docs and 1–3 weeks for scans/forms; this pipeline returns in minutes, powered by their model Chandra on the same harness as the JATS pipeline.
- How to use: available via API (outputshape=wcag) or in the playground, checked against automated tests and a real screen reader.
Related event: Datalab Launches PDF Accessibility API, Cutting Cost from $100+ to Cents(4 posts)→
More from Apps
- AI Overviews on 80% of queries will make Google Search Console data a mess — lilyraynyc · 2026-10-27
- Law professor uses Opus 5.5 to produce a Like Company v Google copyright explainer video — technollama · 2026-10-27
- Open-source Clairic Studio turns pasted text into typed data and working code in 5 languages — dfinke · 2026-10-27
- Google moves inline links up in AI Mode, tests surfacing official task pages on top — nikvassev · 2026-10-27
- Cardboard launches as a note-taking companion for ChatGPT and Claude — pritisinghhhh · 2026-10-26
- Non-coder builds 93-stage arcade game entirely with Claude Code, describes dev workflow — TwoSome22 · 2026-10-26