{
  "title": "For AI agents",
  "description": "Local MCP server, llms.txt, and llms-full.txt — everything an LLM or coding agent needs to use nori-ui correctly.",
  "url": "/docs/for-ai",
  "since": "0.1.0",
  "tags": [
    "meta"
  ],
  "platform": "both",
  "source": "---\ntitle: For AI agents\ndescription: Local MCP server, llms.txt, and llms-full.txt — everything an LLM or coding agent needs to use nori-ui correctly.\nsince: 0.1.0\ntags: [meta]\n---\n\nimport { Callout } from 'fumadocs-ui/components/callout';\n\nThe library was built with both human and machine readers in mind. Three surfaces make the docs first-class for LLMs and coding agents:\n\n## Local MCP server (recommended)\n\nInstall [`@nori-ui/mcp`](https://www.npmjs.com/package/@nori-ui/mcp) and let your agent run it locally over stdio. The package bundles the entire docs corpus, so there's no network dependency and no rate limit.\n\n### Claude Desktop / Claude Code\n\nAdd to your MCP config (`~/.claude.json`, `claude_desktop_config.json`, etc.):\n\n```json\n{\n  \"mcpServers\": {\n    \"nori-ui\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@nori-ui/mcp\"]\n    }\n  }\n}\n```\n\n### Cursor\n\n`~/.cursor/mcp.json`:\n\n```json\n{\n  \"mcpServers\": {\n    \"nori-ui\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@nori-ui/mcp\"]\n    }\n  }\n}\n```\n\n### Any stdio-compatible agent\n\nThe CLI speaks the standard MCP stdio transport. Point your client at `npx @nori-ui/mcp` (or install globally and use `nori-ui-mcp` directly).\n\n### Tools exposed\n\n- `search_components(query)` — fuzzy search by name, description, or tag.\n- `get_component_docs(name)` — full docs body for one component.\n- `get_component_props(name)` — typed prop list with descriptions and defaults.\n- `list_examples(name)` — runnable code snippets keyed by component.\n\n<Callout type=\"info\">\nThe data corpus is baked into the published `@nori-ui/mcp` package, so the version your agent installs matches the library version it queries — no silent drift between docs and answers.\n</Callout>\n\n## HTTP MCP endpoint (fallback)\n\nFor browser-based clients that can't spawn a local process — Claude.ai web, ChatGPT custom GPTs, hosted MCP gateways — the same server is reachable over HTTP:\n\n```\nhttps://nori-ui.com/mcp\n```\n\nSame four tools, same data, same answers. Prefer the local server when you have the choice; the HTTP endpoint is a quiet fallback.\n\n## llms.txt\n\nA flat index of every doc page, optimised for token-efficient discovery. The convention follows [llmstxt.org](https://llmstxt.org).\n\n```\nhttps://nori-ui.com/llms.txt\n```\n\n## llms-full.txt\n\nEvery doc page concatenated into a single plain-text file — the entire library in one fetch.\n\n```\nhttps://nori-ui.com/llms-full.txt\n```\n\nUse this when context is cheap and you want the agent to have the whole reference up front. For long-running sessions, prefer the MCP server (local or HTTP): targeted, structured, and avoids token bloat.\n"
}