Buy from an MCP client

@zanora/mcp is an MCP server that runs locally over stdio. It gives the model eleven tools. Two of them matter most: zanora_discover finds something to buy, and zanora_invoke pays for it and calls it.

#Set up in two commands

If you don't have a Zanora account yet, the setup command creates one, gives an agent a wallet, and stores the keys on your computer:

npx -y @zanora/mcp setup --email you@example.com --org "Northwind" --name claude-code --approval-rules
# Zanora emails you a 6-digit code. Then:
npx -y @zanora/mcp setup --code 123456

• You sign up. The account only exists once you enter the code from your inbox. --approval-rules makes purchases over $2.00, and any from an unverified seller, wait for your approval in the console.
• Nothing secret is printed. The agent's signing key and its API key go to ~/.zanora (ZANORA_HOME), readable only by you. The API key is an agent key: it can buy, never approve.
• The client config holds no secret. The server reads ~/.zanora when ZANORA_API_KEY isn't set:

{ "mcpServers": { "zanora": { "command": "npx", "args": ["-y", "@zanora/mcp"], "env": { "ZANORA_MAX_PRICE_MICROS": "5000000" } } } }

In Claude Code: claude mcp add zanora -- npx -y @zanora/mcp. Also: setup --resend sends a new code, setup --status shows what's set up, and --force starts over. An AI agent can run the same commands for you; zanora.dev/skill.md↗ tells it how.

Already have an account and an agent from the console? Use the config below instead.

#Add it to your client

{
  "mcpServers": {
    "zanora": {
      "command": "npx",
      "args": ["-y", "@zanora/mcp"],
      "env": {
        "ZANORA_API_KEY": "zk.akey_…",
        "ZANORA_AGENT_KEY_FILE": "/absolute/path/to/agent-key.pem",
        "ZANORA_MAX_PRICE_MICROS": "5000000"
      }
    }
  }
}

Restart the client. The server checks the configuration at startup and prints a banner to stderr. A misconfiguration fails immediately with the name of the variable at fault.

#Configuration

¹ One of the two, unless setup stored the key file (see ²). ² Unless you ran setup: with ZANORA_API_KEY unset, the server reads the agent key, wallet and key file it stored in ~/.zanora (ZANORA_HOME).

#The tools

The loop is discover → invoke with the capabilityId from a result. The server remembers endpoints from discovery, so the model never types a URL. An id it hasn't seen is an error, not a guess, and zanora_capability resolves it properly. REST and MCP capabilities are both bought with zanora_invoke.

Every tool returns typed structuredContent along with text. zanora_invoke reports four named progress steps if the client sends a progress token.

#When the answer is "no"

Refusals come back as readable tool results, and nothing is charged:

#Resources

The server also offers resources a person can browse without asking the model: zanora://wallet (balance) and zanora://purchases (recent receipts). They return the same data as the matching tools.

#Try it

Ask the model:

Search Zanora for invoice OCR under $2, buy the cheapest one on
https://example.com/invoice.png, verify the receipt, and rate the seller.

#Safety model

• The key file is the spending authority. Give each agent its own wallet, funded with what you'd accept losing to it.
• The limits are on the server. Budgets and policies are checked by the platform. ZANORA_MAX_PRICE_MICROS is an extra limit on your side, not a replacement.
• Buying and reading only. The server can't create wallets, move money between them, change policies or read another tenant's data.
• Running code is off by default. A search result can never cause a seller's code to run on your machine.