For readers, Origami changes the publication from something that is merely displayed into something that can be acted on while remaining tied to the authoritative source. Different views, annotations, AI assistance, spatial arrangements, and computational exploration can all preserve the identity of the exact passage, figure, equation, definition, or reference involved.
• Cite exactly what was read.
• Copy a citation to a paragraph rather than only to a document.
• Return directly to the cited passage.
• Potential restoration of reading context.
• Definitions at the point of reading.
• Multiple glossary presentation modes.
• Concept-first reading.
• Declutter by concept.
• Fold to headings.
• Fold to concepts.
• Fold to citations.
• Fold to names or people.
• Overview reading.
• Focus reading.
• Horizontal or column reading.
• Outline reading.
• Transcript-specific reading views.
• Stretchtext for optional expansion of secondary material.
• Interactive tables.
• Resettable experimentation with authorial defaults preserved.
• Structured mathematics rather than equation images.
• Copy mathematical source where available.
• Future equation exploration for variables, units, and first occurrences.
• Inspect real 3D objects.
• Pull a figure into spatial computing.
• See the author’s intended poster view first.
• Manipulate a 3D object independently.
• View objects at meaningful physical size where known.
• Retrieve original model files.
• Follow model provenance.
• Read the author’s concept map.
• Walk through an argument in XR.
• Compare distant passages simultaneously.
• Lift addressed elements into a workspace without losing source connections.
• Build personal spatial arrangements without modifying the book.
• Highlights and notes remain external to the immutable publication.
• Annotations can anchor to semantic elements rather than page coordinates.
• Portable annotation models become possible.
• Safer annotation migration across editions.
• Exact and inferred reattachment can be distinguished.
• Quote-text fallback for annotation re-anchoring.
• Group several editions as one work.
• Inspect what changed between editions.
• Follow lineage across revisions.
• See superseded material in context.
• Library-wide concept navigation.
• Library-wide citation navigation.
• Citation trees.
• Citation-context views.
• Concept maps across books.
• Timeline and temporal views.
• Author or person views.
• Read-aloud over semantic paragraphs.
• Attention aids over stable structure.
• Personal typography without breaking references.
• Freedom from fixed pages.
• Ask AI about an exact passage.
• AI answers with checkable citations.
• AI explanations using publisher-supplied concepts.
• AI access to canonical bibliography records.
• AI comparison of editions using identity and lineage.
• AI summaries that remain navigable back to source elements.
• AI-generated study guides grounded in source structure.
• ‘Show me the evidence for this claim’ can resolve to actual elements.
• Text-only LLM discovery of richer metadata through the visible colophon.
• Citations can become interactions rather than dead reference markers: a capable reader can offer actions such as viewing bibliographic details, copying a citation, locating a local copy, opening an online rendition, or acquiring the cited work.
• A Copy-to-Cite workflow can create a citation to the exact addressed source element rather than forcing the reader to reconstruct it manually.
• Shareable reading addresses can carry an exact source location and, as an external reader-layer convention, may also carry view, furl/expand state, or selected context without changing the publication.
• External highlights and margin notes can follow the reader across devices and can interoperate with Web Annotation systems.
• A reader can distinguish the author’s published map or spatial arrangement from the reader’s own personal layout.
• A lineage or citation-history view can place the publication in its cites/cited-by context.
• Glossary terms can become live cross-document context objects rather than isolated end-matter definitions.
• A reader can switch between a liquid semantic view and a fixed PDF/print rendition when both are available.
• A private on-device LLM can work over the structured publication and return answers grounded in exact elements and real citations.
• Local-library links can resolve directly between publications while still retaining ordinary bibliographic and Web-addressable representations.