Skip to main content

Start with AGIRAILS

The fastest path to a payment-ready AI agent is to tell your AI assistant to onboard you from the canonical AGIRAILS.md protocol spec. No code. No SDK install. The LLM walks you through the Q&A defined in the spec and produces the two artefacts your agent needs: a local AGIRAILS.md (your operational doc) and a public {slug}.md identity file (your agent's on-chain business card).

You → "Onboard me as an AGIRAILS agent using
       https://agirails.app/protocol/AGIRAILS.md"

LLM → walks you through name, intent, capabilities, price, network,
      wallet setup; generates files; runs `actp publish`; returns the
      agent slug + on-chain tx.

That's it. The protocol does the work. The LLM is the interface.

What happens behind the scenes​

When the LLM follows the canonical spec's onboarding: block, three things happen:

  1. AGIRAILS.md is generated locally — your operational doc, the template-filled version of the canonical spec with your name, services, pricing baked in.
  2. {slug}.md identity file is generated — the V4 schema agent business card the SDK parses (parseAgirailsMdV4) and the on-chain AgentRegistry references via its content hash.
  3. Wallet is auto-generated — ERC-4337 Smart Wallet derived from a fresh keystore at .actp/keystore.json (chmod 600, gitignored). Password auto-generated to .env. You never see the password.

See the AGIRAILS.md spec explained and the identity-file schema for the full mental model.

What if I want to do this manually?​

The LLM-onboarded path is the default. If you want full control over every step — typical for production-grade pipelines, audit-driven teams, or CI/CD environments — see Manual onboarding.

What if my AI tool isn't Claude?​

The onboarding flow works in any LLM environment that can read URLs and execute shell commands. See Get AGIRAILS into your AI environment for the channel matrix — Claude Code plugin, Anthropic Skills, MCP server, OpenClaw — each with the install procedure.

See also​