A reader highlights any descriptive passage in a digital text — five words or more. An AI classifies it, routes it to the appropriate generation module, and produces a contextually appropriate visual inline, without the reader ever leaving the document.
Not a summary. Not a link. A bespoke visual — generated from the text itself, at the moment the reader needs it.
From the prototype
- A cardiac cycle — animated from a Gray's Anatomy passage, ECG trace synchronised at 72 BPM
- A battle formation — tactical map of Waterloo, Allied positions and French advance
- A pharmacological cascade — beta-blocker mechanism, receptor to ion channel
- A voyage route — Pequod traced from Nantucket to the Japan Ground
- A storm system — pressure low reconstructed from the passage, tracking north-east
- A chemical reaction — neutralisation shown as a process, not an equation
Reader-initiated. The visual appears only when the reader chooses.
Nothing is embedded. Nothing interrupts. The feature is invisible until wanted.
No content modification required. The augmentation layer sits within
the reading interface. Existing content files, formats and distribution pipelines
are untouched.
Platform and model agnostic. Works across EPUB, PDF and web-based
reading interfaces. Implementable using any AI model. No vendor lock-in.
For publishers, platforms, distributors and device manufacturers.
The patented method applies across the full digital reading ecosystem.
See it working
View the interactive demonstration
Intellectual Property
PTIE20260000000265
Short-Term Patent Application · Intellectual Property Office of Ireland
Filed April 2026 · Paris Convention window open to April 2027
Filed April 2026 · Paris Convention window open to April 2027
Contact
Colm O'Connor
colm@aiaugmentedreading.com
Waterford, Ireland
Waterford, Ireland