<aside> 🧭
</aside>
Connecting MCP Client to Cashmere MCP Server
How to access Wiley Content in Perplexity
Preparing Structured Data for Ingestion
Uploading New Content via the website
Cashmere Link Guide (OAuth) Overview
Third-Party OAuth Integration Guide for Publishers (BYOL / Short Link)
Connecting Wiley with Perplexity | DEPRECATED
Cashmere supports fine-grained licensing controls that determine how licensed applications can use, retrieve, and display publisher content. Two primary rights govern usage in AI and RAG-based workflows:
RAGREAD_RAG_SOURCEFor all RAG applications, RAG is a required right. However, it can be granted separately or together with READ_RAG_SOURCE, which requires a consumer to have the RAG license right. This allows publishers to define whether a licensee may use content in a RAG pipeline, whether they may see the underlying source, or both.
RAG RightRAG Right PermitsThe RAG right grants a licensee permission to use licensed content in Retrieval-Augmented Generation (RAG) applications.
In practice on Cashmere, this includes:
Ability to call search_publications in the MCP server
Licensees can submit queries against the publisher’s Omnipubs, retrieve relevant snippets, and use those snippets inside their AI workflows.
Ability to integrate the MCP Server into their RAG pipeline
The licensee may treat the Cashmere MCP Server as part of their retrieval infrastructure.
The RAG right does not grant access to full source content—the licensee is only granted access to chunks of content with source data that can be referenced in their LLM queries.
Licensees may retrieve and use snippets returned by search, but they cannot see full sections, chapters, or documents unless they also hold READ_RAG_SOURCE.
READ_RAG_SOURCE RightThe READ_RAG_SOURCE right grants the licensee access to the full source context of content used in a RAG workflow.
When granted:
This right is important for use cases where:

Sample content page when the licensee has READ_RAG_SOURCE license right.
READ_RAG_SOURCE Right Does Not PermitThis right does not automatically allow:
Those uses typically require separate rights negotiated with the publisher.
READ_RAG_SOURCEIf a licensee has RAG but not READ_RAG_SOURCE, access is limited to what we call the Landing Page Experience.
When a snippet is returned:

Sample landing page for licensee that does NOT have the READ_RAG_SOURCE license right
Publishers control where users go next. Examples:
Redirect to a purchase or checkout page
(e.g., “Buy the book to read the full chapter”)
Redirect to the publisher’s online reading platform
Redirect to a signup or subscription flow
The landing page acts as a neutral bridge: it shows enough context for the user to confirm relevance but still respects the publisher’s monetization strategy.
Cashmere supports automatic redirects without showing a landing page, but we strongly recommend keeping landing pages enabled.
Through repeated testing, we’ve learned:
Landing pages give just enough visibility for the user to confirm:
This is particularly important in RAG workflows where the user may be exploring or iterating on a question.
| Right | Allows Use in RAG Workflow | Allows Viewing Full Source | User Experience Without the Right |
|---|---|---|---|
RAG |
Yes | No | Snippet only; landing page; redirect to publisher |
READ_RAG_SOURCE |
Not required for basic RAG, but enables richer workflows | Yes | Without this: limited preview + timed link + controlled redirect |
| Both | Full RAG integration | Full context viewing | Best experience for advanced AI use cases |