Idea: optional Next.js recipe for AI-readable pages from the same MDX collection

[mustcanbedo]

Hi! I maintain next-ai-ready, a Next.js package that generates machine-readable documentation from MD/MDX.

I would like to discuss a small, opt-in Velite recipe: Velite would continue to own the application's content data, while next-ai-ready reads the same source collection for llms.txt and page Markdown. It would not replace Velite's loaders or change the existing page UI.

The useful part to test would be:

• Map a small allowlist of published MDX pages to the same Next.js routes used by the Velite collection.
• Preserve canonical page URLs and exclude drafts from the machine-readable output.
• Document build ordering without adding the SDK to client bundles.

There is an existing Nextra example showing the shared-source approach, but I have not validated a Velite integration yet:
https://github.com/mustcanbedo/next-ai-ready/tree/main/examples/nextra-docs

Would this be useful as an optional example or community recipe? In particular, is there an existing route/slug mapping convention or similar recipe I should reuse? I would like to agree on the smallest useful scope before submitting an implementation PR.