Complete book metadata — extracted, enriched, and verified
Most tools can only reshape the metadata you already have. Origami reads the actual EPUB or PDF, fills every gap with AI, and checks the result against the book’s own text — one clean, ONIX-ready record with every field sourced.
Start from the book, not a feed
Most metadata automation starts from the ONIX you already have and makes it look better. It can’t tell you the declared language is wrong, that an ISBN doesn’t belong, or that an obvious subject is missing — because it never opens the book.
Origami does. It reads the EPUB or PDF itself, so enrichment is grounded in what’s actually inside, and every value carries its source: taken from the file, inferred by AI, or verified against the text. That’s the difference between a tidier spreadsheet and a record you can trust.
- ISBN-13
- 978-84-1234-567-8 From the file
- Language
- Spanish Verified
- Contributors
- 2 · with roles AI
- BISAC
- FIC019000 AI
- Thema
- FBA AI
- Keywords
- 24 terms AI
- Accessibility
- WCAG 2.1 AA Verified
- Description
- short + long Your edit
One pass, the whole record
Extraction, enrichment, verification and standards in a single run — not four tools stitched together.
-
01
Extract
Read every field already embedded in the EPUB or PDF — no re-keying.
-
02
Enrich
Fill the gaps with AI: subjects, keywords, contributor roles and descriptions.
-
03
Verify
Cross-check the record against the book’s own text and flag what doesn’t match.
-
04
Standardize
Normalize everything to ONIX for Books 3 with BISAC and Thema subjects.
-
05
Publish
Export the record, feed retailers and libraries, or write it back into the file.
Everything a record needs
Not a handful of fields — the complete bibliographic record the trade expects, in one place.
Identifiers
ISBN-13 and ISBN-10, product form and edition — correctly typed to the ONIX identifier lists.
Contributors & roles
Authors, editors, translators and illustrators, each mapped to an ONIX List 17 role code.
Titles & imprint
Title, subtitle, publisher and imprint — cleaned and consistently cased.
BISAC & Thema
Subject codes in both the North-American (BISAC) and international (Thema) schemes.
Keywords
Retail keywords in the language of the book, generated for discovery — not filler.
Descriptions
Short and long descriptions that follow the ONIX text-type conventions retailers expect.
Language
The declared language, checked against the prose and corrected when they disagree.
Accessibility
ONIX accessibility features (List 196) derived from the file — the metadata the EAA needs.
Prices & availability
For publica.la titles, live price and availability folded straight into the record.
Every field is sourced — and your data always wins
Enrichment you can audit. Each value is tagged with where it came from, and nothing you decide is ever silently overwritten.
Read from the file
Whatever the EPUB or PDF already declares is taken verbatim and marked as embedded.
Enriched by AI
Missing fields — subjects, keywords, roles — are inferred in the language of the book and clearly labelled.
Verified against the text
The declared language, ISBN and key subjects are checked against what the book actually contains.
Edited by you, and kept
Any value you change is locked. Re-running enrichment never overwrites your edits.
Feed-only automation vs Origami
Both hand you cleaner ONIX. Only one of them actually reads the book.
| Capability | Feed-only tools | Origami |
|---|---|---|
| Reads the actual book file (EPUB / PDF) | — | |
| Fills genuinely missing fields with AI | Partial | |
| Verifies metadata against the text | — | |
| Shows the source of every field | — | |
| Your embedded data always wins | — | |
| ONIX for Books 3 output | ||
| BISAC & Thema subjects | ||
| Accessibility metadata (EAA) | — | |
| Writes corrections back into the file | — |
Built on the standards the trade runs on
The record leaves Origami ready for retailers, aggregators and libraries — no re-mapping.
ONIX for Books 3
Every record is normalized to ONIX 3.0 (3.1-ready), the global standard maintained by EDItEUR.
BISAC & Thema
Subjects in both schemes, so a title is shelved correctly in every market.
EDItEUR code lists
Roles, languages, countries and product forms drawn from the official code lists (Issue 73).
Accessibility (EAA)
ONIX List 196 accessibility metadata — required to sell ebooks in the EU under the European Accessibility Act.
How it works
Upload a file
Drop in an EPUB or PDF from your workspace.
We read, enrich and verify
Origami reads what’s embedded, fills the gaps with AI, and checks it against the text.
Export or write back
Review the sourced record, feed it to your channels, or write it into the file.
Questions about metadata enrichment
- How is this different from a metadata feed tool?
- Feed tools reshape metadata you already have. Origami reads the book itself, so it can fill genuinely missing fields and verify the rest against the text — not just re-format a spreadsheet.
- How is this different from Origami’s metadata extraction?
- Extraction pulls what’s already in the file. Enrichment goes further — it fills the missing fields with AI, verifies them against the text, and completes the ONIX record.
- What formats can it read?
- EPUB and PDF.
- Does AI overwrite the metadata in my file?
- No. Anything embedded in the file always wins, and any value you edit is locked — AI only fills what is genuinely missing.
- What does “verify against the text” mean?
- Origami compares the declared metadata with the book’s own content — flagging, for example, a language tag that doesn’t match the prose or an identifier that doesn’t belong — so errors surface before they reach retailers.
- Will it produce accessibility metadata for the EAA?
- Yes. Origami derives ONIX List 196 accessibility features from the file — the metadata the European Accessibility Act requires to sell ebooks in the EU.
- Is my content used to train AI?
- No. Origami’s AI providers do not use your content to train models — see our AI policy for the details.
Metadata is normalized to the ONIX for Books 3 standard maintained by EDItEUR, using code lists Issue 73. BISAC is a trademark of BISG; Thema is maintained by EDItEUR. Accessibility features follow ONIX List 196. AI-inferred fields are drafts — review before publishing.
Enrich your catalogue’s metadata in Origami
Start free and build a complete, verified record for your first title.