name: Agent-Ready Docs
description: >-
  Capability composition for delivering AI-era documentation. Combines llms.txt
  generation, the Fern MCP server, Ask Fern AI search, and Fern's agent-score
  audit to make every page consumable by LLMs and coding agents like Claude
  Code, Cursor, and Windsurf.
provider: Fern
providerId: fern
created: '2026-05-22'
tags:
  - Agent Ready
  - AI
  - Documentation
  - MCP
  - llms.txt
workflow:
  - step: enable-llms-txt
    capability: fern.docs.llmsTxt
    description: Turn on llms.txt generation for the docs site.
    inputs:
      - hostedDocsSite
    outputs:
      - llmsTxtArtifact
  - step: enable-mcp-server
    capability: fern.mcp.askFern
    description: Register the docs with the Fern MCP server so MCP clients can call ask_fern_ai.
    inputs:
      - hostedDocsSite
    outputs:
      - mcpEndpoint
  - step: enable-ask-fern
    description: Enable Ask Fern in-docs AI search grounded in the customer's docs, SDKs, and guides.
    inputs:
      - hostedDocsSite
    outputs:
      - askFernEnabled
  - step: audit-agent-readiness
    capability: fern.docs.agentScore
    description: Run Fern's agent-score tool to quantify how navigable and parseable the docs are for AI agents.
    inputs:
      - hostedDocsSite
    outputs:
      - agentReadinessReport
  - step: validate-against-clients
    description: Run smoke tests against Claude Desktop, Claude Code, Cursor, and Windsurf using the MCP endpoint and llms.txt.
    inputs:
      - mcpEndpoint
      - llmsTxtArtifact
    outputs:
      - clientCompatibilityReport