Every book listing you have ever seen on a retailer's site was built, at some point, from an ONIX file. ONIX (ONline Information eXchange) is the book trade's agreed format for describing a product: an XML document that carries the title, contributors, description, subjects, price, availability and dozens of optional details.
Who sends it
Large publishers send ONIX feeds directly to retailers and wholesalers. Distributors send feeds on behalf of the publishers they carry. Platforms like KDP and IngramSpark generate ONIX from the forms you fill in. Aggregator databases, Books in Print among them, both receive and redistribute it.
For an independent author, ONIX is usually invisible: you fill in a form and the platform does the rest. It becomes visible the moment you deal with a distributor or retailer that says "send us your ONIX", or when you try to work out why a listing shows the wrong subtitle.
What is in a record
The current version is ONIX 3.0 (3.1 is a minor update). A product record contains blocks such as:
- ProductIdentifier: the ISBN (type 15), and optionally other identifiers.
- DescriptiveDetail: form (paperback is
BC, hardcoverBB, ebookED), title, contributors, language, subjects, extent. - CollateralDetail: the description, review quotes, cover image link.
- PublishingDetail: publisher, imprint, publication date, status.
- ProductSupply: who supplies it, availability, and price by currency.
Each value uses a code from EDItEUR's code lists, which is why ONIX feels bureaucratic: 01 for a main title, A01 for an author, 03 for a description. The codes make the file machine-readable in any language.
Why listings drift
Every destination holds its own copy of your record, updated only when it next receives a feed. If a description was improved in one platform's form, the platforms that receive feeds from elsewhere never learn of it. The cure is to send all destinations from one source.
Getting an ONIX file for your book
If a distributor or retailer asks for ONIX, you have three options: ask the platform that holds your record to export it (not all can), hire a metadata service, or generate it from your own record.
ScanSpine generates ONIX 3.0 for every title from the record you maintain, as a download for one book and as a public feed of all your active titles that a partner can pull on a schedule. Edit the record and the feed changes with it, which is the whole point.
What ONIX will not do
ONIX moves data; it does not decide who accepts it. A retailer still chooses whether to list your book, and a feed is only ingested where there is an agreement to ingest it. That is why trade and library databases appear as "planned" rather than "live" on our distribution list until the agreements exist: a format is not a channel.