MCP · for coding agents

Empower agents with auth and payments.

MCP tools to automate auth and pay — your keys and card never leave the vault.

Read the docs →
Works inside
Claude Code
Codex
Goose
Cursor
OpenCode
claude code — trusty-squire
$add Google OAuth in one prompt
squire opening Google Cloud Console · OAuth client, secret…
secret sealed → vault cred_oauth_••••••
# used via the proxy — never shown to the agent
What developers ask

What is Trusty Squire?

Trusty Squire is an MCP server for coding agents. Claude Code, Codex, Cursor, OpenCode, or Goose plans the job; Trusty Squire operates the website, wires up the integration, or completes the purchase — keeping the generated secret or card on the safe side of the boundary.

How does it work?

  1. Your agent names the outcome — an account, an integration, or a purchase.
  2. Trusty Squire drives the real signup, setup, or checkout flow.
  3. The credential or card is encrypted in a write-only vault and injected server-side when used.

If a site requires a phone, hard CAPTCHA, or a decision only you should make, the run stops and tells you instead of guessing.

Magic happens when your agents can sign up, provision, and purchase on your behalf — hard tasks in one prompt.

The card is never exposed to the agent, and no charge happens without one human approval. API keys stay vaulted — the raw secret never enters the agent’s context.

Install the squire. Let it handle the rest.

One command. Plugs into your agent of choice. Free to start.

Get started →