mcpField
A
Authoryze

Authoryze

Remote
authoryze.aiFinance0Updated 10w ago

About Authoryze

Authoryze is a remote MCP server for agents that need to make purchases under owner-defined merchant and amount limits. Each request can be approved or denied before a single-use virtual card is issued; it supports US-issued cards.

Key features

  1. 01Merchant- and amount-based spending limits
  2. 02Purchase approval and denial flow
  3. 03Single-use virtual card retrieval
  4. 04Spending summaries by time period
  5. 05Purchase request status lookup

Use cases

  1. 01Let an agent buy from specified merchants
  2. 02Review purchase requests before card issuance
  3. 03Track spending against configured limits

MCP support

Authoryze exposes purchase requests, status checks, spending summaries, and approved card retrieval to MCP clients. It supports OAuth 2.1 with PKCE or a static agent API key.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://authoryze.ai/api/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

All endpoints
Streamable HTTPhttps://authoryze.ai/api/mcpAuth required

Tools

Details

From the project README

Authoryze MCP Server

Authoryze gives your AI agent a spending limit. It lets you hand an agent a payment method without handing it your card: you define what it can buy, from which merchants, and up to what amount, and every purchase request comes to you for approval before a single-use virtual card is issued. It is built on Visa Intelligent Commerce and Mastercard Agent Pay, and is currently available for US-issued cards. Authoryze is a remote MCP (Model Context Protocol) server that works with Claude, Claude Code, ChatGPT, Cursor, Windsurf, and any MCP-compatible client.

Installation

OAuth-capable clients such as Claude and ChatGPT need only the endpoint:

{
  "mcpServers": {
    "authoryze": {
      "url": "https://authoryze.ai/api/mcp"
    }
  }
}

Clients that don't support OAuth, such as Claude Code, Codex, and custom frameworks, authenticate with a static agent API key as a bearer token:

{
  "mcpServers": {
    "authoryze": {
      "url": "https://authoryze.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Tools

  • request_purchase - Ask to make a purchase from a specific merchant, for a specific amount, with a reason. The request either gets approved automatically, gets sent to the account owner for approval, gets denied with a reason, or fails in a way that's safe to retry.
  • check_status - Look up what happened to a purchase request: approved, denied, still waiting on approval, or failed. Doesn't reveal any card details itself.
  • get_spending_summary - See how much has been spent so far against the configured limits (daily, weekly, monthly, and total), including any account-wide limit that applies across all of an account's agents.
  • retrieve_card - Once a purchase is approved, get the actual card number, expiry, and security code to complete the purchase. This can only be done once per approved purchase; the details are shown a single time and can't be retrieved again.

Requirements

Authoryze is available for US-issued cards only. International support is pending issuer coverage on the agentic token programs.

Authentication requires either OAuth 2.1 with PKCE (S256) and Dynamic Client Registration, or a static agent API key used as a bearer token.

Links

Homepage: https://authoryze.ai

Docs: https://authoryze.ai/docs

Smithery

Listed on Smithery: https://smithery.ai/servers/authoryze/authoryze

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.