The capability layer for the agentic web

Make any website agent-ready

AgentReady turns a public website into an AI-readable interface for agents: MCP tools, llms.txt, cited answers, capability manifests, and safe action plans. Try it live — no signup.

Try it live

Want to publish your own site? Make it agent-ready in 60 seconds →·Preview a grounded plan →

Distribution layer

One MCP server for any website and AI agent.

AgentReady gives AI agents one standards-based entry point to discover websites, inspect capabilities, ask cited questions, request grounded plans, and keep indexes fresh. Connect any MCP-compatible client without locking the product to one model.

MCP endpoint

https://www.agentready.it.com/api/mcp

Local bridge for any MCP client

npx @agentreadyweb/mcp

CLI — grade any site's agent-readiness

npx @agentreadyweb/mcp grade yourdocs.com

CLI — ask, index, refresh

npx @agentreadyweb/mcp ask yourdocs.com "What does this site do?"npx @agentreadyweb/mcp index https://yourdocs.comnpx @agentreadyweb/mcp refresh yourdocs.com

Use refresh in CI after a docs deploy to keep agent answers current.

No API key. No account. Full setup guide for every client →

Seven MCP tools

  • ask_sitecited answers from any site
  • list_sitesbrowse the index
  • get_site_capabilitiesinspect a site manifest
  • plan_site_actioncreate a grounded plan
  • submit_siteindex a new site in ~60s
  • refresh_siteforce a re-crawl
  • rate_answersend quality feedback

MCP tools cover discovery, grounded Q&A, capability manifests, planning, indexing, freshness, and feedback. The same workflows are available from the CLI: grade, ask, index, and refresh. Permissioned customer actions remain behind explicit confirmation and connector controls.

Make your website AI-readable with llms.txt and live agent endpoints

Paste your URL below. AgentReady crawls your site and creates a ready-to-host llms.txt that follows the llmstxt.org spec — plus live /ask and /mcp endpoints for agents. What is llms.txt? →

How it works

From website to AI agent interface

01

Publish a capability layer

Enter any public website URL. AgentReady discovers and structures your content without requiring a code change.

02

Agents discover and ask

Agents find your manifest and MCP endpoint, then receive grounded answers with citations instead of guessing from page markup.

03

Plans make actions safe

When a request could change something, AgentReady creates a bounded plan that can be reviewed before execution.

What you get

The layers AI agents need

AgentReady starts with reliable knowledge access and adds permissioned execution without locking you to one model or client.

Readable

Live

Clean, structured text extracted from every page — no nav, no ads, no noise. Ready for any LLM to consume.

Findable

Live

A spec-compliant llms.txt index file at your site root. AI agents discover it automatically, just like robots.txt.

Askable

Live

A hosted /ask RAG endpoint and /mcp MCP server so any AI agent can query your site in natural language and get cited answers.

Actionable

Sandbox

Typed, permissioned actions can be previewed in a safe sandbox before a customer connector is enabled.

Why AgentReady

A practical agent layer for every site

AgentReadyScrunchNLWeb
Self-serve
No developer needed
llms.txt generation
Hosted /ask + /mcpSelf-hosted
SMB / long-tail
PriceFree to try · $500 pilot$250+/moFree (DIY)

Comparing AgentReady vs Kapa or GitBook? See the full breakdown →

Ready to connect your systems?

Private indexing, custom capabilities, higher limits, or a pilot connector — tell us what you want agents to do and we'll help you ship it.

Make your website agent-ready

No account. No credit card. Paste a URL and get an llms.txt, MCP server, and cited /ask endpoint in under a minute.

No account needed · Works on any public website · Free to try