Make your site readable by AI search engines and callable by AI agents. One config file, zero changes to your UI.
Your existing Next.js app gains a complete AI interface — discovery, retrieval, and tool execution.
Your content appears in ChatGPT, Perplexity, Google AI Overviews. Structured llms.txt and per-page Markdown make you citable.
Define actions with type-safe schemas. They become OpenAPI endpoints and MCP tools that any agent can invoke.
Your MDX pages compile into a rich knowledge graph. AI systems get summaries, topics, Q&A pairs, and relationships.
First-class MCP server support. Claude Desktop, Cursor, and any MCP client can connect to your site directly.
Open standards only — OpenAPI, JSON-LD, llms.txt, MCP. Works with any AI provider. MIT licensed.
init, build, doctor CLI. Type-safe config. CI-friendly checks. Hot reload in development. Works with any Next.js app.
MDX pages are compiled into a semantic graph. AI search engines get structured, citable content via llms.txt, Markdown, and JSON-LD.
# Acme Documentation
> The complete toolkit for building
> modern web applications.
## Docs
- [Getting Started](/docs/start)
Install and configure in 60 seconds.
- [Authentication](/docs/auth)
Secure your app with built-in auth.
- [Deployment](/docs/deploy)
Deploy anywhere with one command.Define actions with Zod schemas. They become OpenAPI endpoints, MCP tools, and tools.json manifests — all from one definition.
import { defineAction } from "next-ai-ready"
import { z } from "zod"
defineAction({
name: "search_docs",
description: "Search documentation.",
input: z.object({
query: z.string(),
limit: z.number().default(5),
}),
public: true,
handler: async ({ query, limit }) => {
return await searchIndex(query, limit)
},
})