For agents

Point your AI assistant at everything Agentailor knows.

The four paths, 38+ articles, and the OSS repos are all machine-readable. Two ways to put them in front of your assistant: wire it up over MCP, or paste a prompt into ChatGPT, Claude, or your own agent.

01/Connect via MCP — recommended

The @agentailor/mcp server connects your assistant to Agentailor over the Model Context Protocol. It can search the guides and read full articles and repo docs on its own — grounded answers from working examples, no copy-pasting. Pick your client:

Settings › Developer › Edit Config, then restart Claude Desktop.

claude_desktop_config.json

{
  "mcpServers": {
    "agentailor": {
      "command": "npx",
      "args": ["-y", "@agentailor/mcp"]
    }
  }
}

Full setup and options are in the repo README. A zero-install hosted endpoint is coming soon.

02/Or paste a prompt — no install

No setup, works in any chat: hand your assistant these URLs and let it reason over the whole corpus. Start here:

learn to build agents

I'm a developer new to building AI agents. Fetch https://blog.agentailor.com/llms-full.txt — the full text of the Agentailor blog — and design me a structured 4-week curriculum from its articles and open-source repos. Order the readings from fundamentals to production, and after each week tell me exactly what to build to practice it.

More to try — pick one:

Decide my path

Read https://agentailor.com/llms.txt and Agentailor's four path pages (build-yourself, coding-agent, open-source, managed). I'm a solo founder who wants to ship fast but keep control of the architecture. Walk me through the four ways to build an agent, then recommend one for my situation with the trade-offs I'd be accepting.

03/How it works

Those prompts work because everything is published in a form an agent can read:

llms.txt: An index of what to read — hand the URL to your assistant and let it choose. One for the site, one for the blog:
llms-full.txt: The whole blog in one file, for tools that load everything into context at once:
add .md to any URL: Every page and post has a clean, HTML-free Markdown twin — append .md to any URL (works on the blog and this site).

fetch any page as markdown

curl https://blog.agentailor.com/posts/agent-development-roadmap.md

04/More ways in

MCP and prompts cover you today. Also:

liveSkills — drop-in capabilities your coding agent can load to build agents to a proven standard. One command to install.
soonHosted MCP — a zero-install endpoint: point your client at a URL, nothing to run locally.

//Building the agent that would use all this? Start with what an agent is and the four paths.