Home Reports Technology How It Works Demo Download App

Narrative enrichment infrastructure

Turn long-form fiction into a spoiler-safe reading experience.

EnrichReader transforms a reader's progress into the right context: evolving character profiles, connected lore, and traceable narrative facts that never reveal the next chapter too soon.

Explore the reader See public reports

Working product

Reader app available on web and Android

Demonstrated output

Public narrative reports across multiple long-form series

Local-first reader

Reading and spoiler controls stay on the device

The product

A living layer for a book a reader already owns.

01

Progressive knowledge

Profiles and lore change with the reader’s actual place in the text instead of exposing the whole series at once.

02

Connected context

Characters, places, organizations, objects, and events become an explorable layer attached to the story.

03

Traceable output

Structured facts are tied to chapter context so the product can preserve narrative order and support verification.

The technology

AI-assisted extraction, with deterministic guardrails.

The enrichment pipeline is designed to make narrative information useful without treating a model response as the final product.

  1. 1

    Structure the text locally

    Book content is organized into chapters and blocks, while local named-entity detection identifies the people, places, organizations, objects, and events worth tracking.

  2. 2

    Use Gemini through Vertex AI for narrative interpretation

    Gemini supports ambiguous entity classification, chapter-context fact extraction, and contextual identity resolution as the Skin-generation pipeline builds the structured reading layer.

  3. 3

    Verify before packaging

    Deterministic checks validate evidence-linked fields, maintain narrative invariants, and version facts by chapter before the final Skin is packaged for the reader.

  4. 4

    Render a local-first reader experience

    The reader combines a local book with the resulting Skin to reveal the right context at the right point in the story, including when the reader is offline.

Public proof

Explore the narrative data.

Our public reports show the kind of structured output the pipeline can produce: entity journeys, mention patterns, narrative gaps, and more.

Browse public reports

Report preview

Lord of the Mysteries

Explore a public narrative analysis built from structured entities and chapter-level data.

Open report →

Build with EnrichReader

Follow the work as we build the next generation of reading tools.

Try the product, explore the public research, or contact the founder to discuss the technology.

Try EnrichReader Contact the founder