Valyu for publishers
Full-stack AI distribution for your archive and catalogue.
We enrich every format, including tables, figures and images, distribute it to AI agents through Valyu or your own channels, and show you how it's used.
What the AI agent gets
As the port strike continues, freight rates rose for a third week, leaving Asia to Europe rates up by a fifth.
- Wire story Port strike lifts shipping costs
- Market report Container freight outlook
Enrichment
Every format, made ready for AI agents.
Archives, backlists and catalogues. Tables, figures and images included. No engineering or maintenance.
Journal articles and market reports
XML
Wire stories and news feeds
EPUB
Books and backlist
Scanned documents
Magazine back issues
Image archives
News and archive photographs
Freight rates rose for a third week as talks stalled…
| Route | Week |
|---|---|
| Asia to Europe | +12% |
| Transpacific | +3% |
photo: concrete lattice facade
Distribution
Two ways to reach AI agents.
Use one or both. Both run alongside your existing licensing and paywalls.
Through Valyu
A source for AI agents and apps built on Valyu, and for assistants such as Claude and ChatGPT through Valyu's connectors.
Revenue shareThe most competitive available. Paid per retrieval, monthly above a threshold.
On your own channels
AI distribution infrastructure for your own channels. No engineering or maintenance on your side.
TermsAgreed per deal.
Programmatic permissioning
How it works
From archive to cited passage.
Set up once. Then every time an agent uses your content, it is metered and attributed.
Set up once Valyu does this, with no engineering on your side.
Every use An AI agent sends a query.
Analytics
See how AI agents use your catalogue.
Retrieval volumes, top documents, query patterns and the applications using your content.
AnalyticsTop documents
RetrievalsProof
The search behind your catalogue.
72.7%
#1 on DRACO, the deep-research benchmark, at under half the cost of the next best.
#1
On ScholarQA, the benchmark for research across the scholarly literature.
SOC 2
Type II, independently audited.
Questions
Questions publishers ask.
Which formats can you work with?
PDF, XML, EPUB, scanned documents and image archives. We extract tables, label images and add metadata, so agents can use more than the text.
What do agents actually receive?
The passages relevant to their task, whether text, a table or a figure, each with a citation to your source. Not your whole catalogue.
How does our content reach agents?
Through Valyu: the Valyu Search and Deep Research APIs, and Valyu's connectors for assistants such as Claude and ChatGPT. On your own channels: search indices, API endpoints and MCP servers that Valyu builds and maintains for you.
Can AI companies train models on our content?
No. Agents never get your catalogue, only the passages relevant to a task, at the moment they need them, each cited and logged. We monitor usage, rate-limit access to stop bulk extraction, and apply programmatic permissioning. On top of that, Valyu's customer terms forbid training on it, storing or caching it, showing its full text to end users, or reselling it.
How does the revenue share work?
Through Valyu, you earn a revenue share on every retrieval, the most competitive available, paid monthly above a threshold. On your own channels, terms are agreed per deal.
Does this replace our existing licensing?
No. It runs alongside your existing licences and paywalls, and there's no engineering or maintenance on your side.
Talk to us
Talk to us about your catalogue.
We'll walk through your archive and the routes that fit it.