The development edit your manuscripts need, in a fraction of the time, with the judgment till yours

Lush makes AI tools for book and educational publishers. They suggest; your editors decide. Copyright-safe AI.

Book a demo
Editorial AI

Edith

More editorial capacity. Built around your standards.

Product features backed by peer-reviewed research from the people who build Lush

Your briefs, your house style, your editorial judgment. Edith brings them together in an AI workspace built for the demands of real publishing. From substantive editing to the detail of language and consistency, it helps your team move manuscripts forward with less manual work and more time for editorial decisions. Every suggested change is yours to review. Every final call stays with you.

Built in our own editorial office. Shaped by a year of work on client manuscripts.

What comes back
Author queries
What's wrong, what it should be, and why, tied to the exact passage, figure, table, or example
Technical and logic checks
Arithmetic, reconciliations, answers checked against their questions, internal consistency across the whole work
Tracked-changes manuscript
Language, style, and consistency edits marked on the file
Editorial report
A title-level read of what works, what breaks, where, and what needs the author
A11y

Access

Accessibility for backlists, archives, and courseware

Access turns PDFs, image-heavy chapters, tables, and course materials into accessible files, with image descriptions that actually describe the image. Built for the EAA and ADA backlog.

What comes back
Accessible files
Tagged PDF and EPUB that hold up to WCAG review
Image descriptions
Real descriptions of each figure, chart, and table, not the filename
Whole books, fast
Extraction, description, and tagging across the full title
Documents it takes
  • Book PDFs
  • Scanned pages
  • Image-heavy chapters
  • Complex image descriptions
  • Tables
Courseware it takes
  • Slides
  • LMS modules
  • Worksheets
  • Assessments
  • Video
  • Audio
  • Learning objects
Why Lush

Runs on your systems

Your manuscripts never leave your network. No third-party AI, nothing sent off to train someone else's model.

It suggests; you decide

Every edit is a tracked suggestion an editor signs off. The AI never gets the final word.

It catches what's slow to catch by hand

An answer that contradicts its own question, a total that won't reconcile, a worked example that doesn't compute, a price rise that never got applied. The substantive checks a subject expert runs, across the whole book.

Built by editors

We run an editorial office ourselves. We built this because we needed it.

Research

Published work

Peer-reviewed writing from the people who build Lush

The invisible editor: operationalizing ‘meaningful human oversight’ in artificial intelligence–mediated workflows

Syed Mohamad Shajahan S M, Founder and CEO, Lush Transnational · European Science Editing 52:e186943 · September 2026 · Open access, CC BY 4.0

When a publisher’s copyediting or peer-review triage runs on an undisclosed AI system, who is responsible for the result? This viewpoint names the risk, semantic drift: edits that keep the grammar and change the meaning. It proposes a five-level disclosure protocol that matches human oversight to what the AI did, from spellcheck to decision support.

Read the paper

Ready to see it on your own manuscripts?

Book a demo

Deployed in production · Built by editors and publishing professionals