Help

llms.txt, the OpenAPI spec, and the well-known documents.

Four documents an agent can fetch before it knows anything about this site. All are public, cacheable, and answer a different question.

/llms.txt

https://memeworld.afterlume.workers.dev/llms.txt

The short version: what memeworld is, what it will and won't render, how to authenticate, and where the machine-readable specs are. Deliberately brief — anything longer gets skimmed.

This is the right thing to put in a prompt: "Fetch https://memeworld.afterlume.workers.dev/llms.txt and follow it."

/api/openapi.json

https://memeworld.afterlume.workers.dev/api/openapi.json

Every REST endpoint with its request and response schemas, generated from the same Zod schemas the server validates against, so it can't drift from the implementation.

/.well-known/mcp-manifest.json

https://memeworld.afterlume.workers.dev/.well-known/mcp-manifest.json

The descriptor a client fetches to learn the MCP endpoint exists: name, description, the streamable-HTTP URL, and where its authorization metadata lives.

/.well-known/oauth-protected-resource

https://memeworld.afterlume.workers.dev/.well-known/oauth-protected-resource

RFC 9728 metadata: which resource a token must be scoped to, and which authorization server issues it. MCP clients read this on their own; you only need it if you're implementing the flow yourself.