How to Connect Mercury to Gemini via MCP

💡 No install? Use cloud.anythingmcp.com directly. Sign in, click Connectors → Mercury, paste your credentials, mint an MCP API key and you're done. No Docker, no git clone, no local server.

Mercury + Gemini

Let your AI read your Mercury business banking: accounts and balances, transactions, cards, statements, recipients, Treasury, credit and users. Works with a Read Only API token and cannot send money. Card freeze and transaction note actions are optional and off by default.

Prerequisites

See the full setup instructions baked into the connector (visible in the in-app store when you select the connector). The required environment variables for this connector are:

MERCURY_API_TOKEN

Step 1 — Get credentials

It cannot move money, by design. The connector has no tool for payments, transfers between accounts, payment requests, new recipients, invoices or new cards, not even a switched-off one, and it never calls Mercury's send-money endpoints. It can list payment requests waiting for approval, but never create, approve or cancel them.

1. Sign in to https://app.mercury.com, click your organization at the top left, then Settings, then Tokens. Your user must be allowed to create API tokens; ask an admin otherwise.
2. Click Create an API Token and choose Read Only. A Read Only token reads all data, cannot move money and needs no IP allow-list.
3. Copy the token into MERCURY_API_TOKEN. It looks like secret-token:mercury_production_...; paste the whole value, including secret-token:. Mercury shows it only once.
4. Install the connector and check it with mercury_get_organization: it shows your legal business name and plan.

Never use a Read and Write token: it can send money without approval. With a Read Only token, Mercury refuses every write.

Switched off at install: mercury_freeze_card, mercury_unfreeze_card and mercury_update_transaction (sets or clears a transaction's note and custom category and overwrites the old values). They need a Custom token with only the matching write permissions, and Mercury requires an IP allow-list for any write permission; on AnythingMCP Cloud, ask support for the current outgoing IP address first. The AI then shows you the card (nickname and last four digits) or the transaction and the change, and waits for your yes. Mercury downgrades a token that has not used its write permission for 45 days, and deletes tokens that have not been used for 45 days.

Financial data is sensitive: balances, transactions, statements, recipients and employee details are confidential company data. Share the MCP server only with people allowed to see them; the AI reads only what you ask for. Account numbers, routing numbers, IBANs and card numbers are removed from every answer; statements and transaction attachments come as download links instead.

What it reads: your organization; checking and savings accounts with available and current balance; transactions per account (the last 30 days unless you name a start date) or across all accounts, with counterparty, note, category, merchant and card, and attachments such as receipts and check images; debit and credit cards with their spend limits; monthly account statements as PDF links; saved payment recipients, without bank details; Treasury accounts with returns, transactions and statements, tax documents included; credit (IO card) accounts; users and their roles; custom expense categories; and payment requests with their approval status.

Amounts: decimal numbers in dollars, negative for money going out.

Sandbox: create a sandbox at https://sandbox.mercury.com/signup, make a token there and change the connector's base URL to https://api-sandbox.mercury.com/api/v1. Production tokens do not work in the sandbox.

Common errors: 401 noTokenInDB: the token is wrong, was deleted after 45 days without use, or belongs to the other environment; create a new Read Only token. 403: the token type or permissions do not allow this call, or a write came from an IP address that is not on the token's allow-list. 404: the id does not exist. 429: wait a moment.

The connector does not send money, create payment requests, recipients, invoices or customers, create or cancel cards, change spend limits, manage webhooks or SAFEs, or upload attachments.

Step 2 — Install the adapter

curl -fsSL https://raw.githubusercontent.com/HelpCode-ai/anythingmcp/main/docker-compose.quickstart.yml -o docker-compose.yml
printf 'JWT_SECRET=%s\nENCRYPTION_KEY=%s\n' "$(openssl rand -hex 32)" "$(openssl rand -hex 32)" > .env
docker compose up -d

Step 3 — Add the connector in Gemini

Your server URL: in AnythingMCP, open MCP Servers → the server this connector is on, and copy its URL (https://cloud.anythingmcp.com/mcp/…). Use it wherever this guide shows YOUR_SERVER_ID.

Gemini CLI reads MCP servers from ~/.gemini/settings.json (on Windows %USERPROFILE%\.gemini\settings.json). Add:

{
  "mcpServers": {
    "anythingmcp": {
      "httpUrl": "https://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID",
      "headers": { "X-API-Key": "YOUR_MCP_API_KEY" }
    }
  }
}

1. Create an MCP API key in AnythingMCP under MCP Servers → your server → API keys, and paste it in place of YOUR_MCP_API_KEY.
2. Save the file and restart gemini.
3. Run /mcp inside Gemini CLI (or gemini mcp list in your shell): anythingmcp should show as connected.
4. Start prompting: Gemini CLI can now call the Mercury tools.

Available tools

Example prompts

• "What are the balances of all our Mercury accounts right now, and how much is in Treasury?"
• "List the outgoing transactions over 5,000 USD from the last 30 days with counterparty, note and category."
• "Which payment requests are waiting for approval? Show the requester, recipient, amount and scheduled date."
• "Give me the download links for last quarter's statements of our checking account."
• "Who has access to our Mercury organization, and with which role?"

FAQ

Where does Gemini CLI look for MCP servers? In ~/.gemini/settings.json for your user, or in .gemini/settings.json inside a project. gemini mcp list shows each server and whether it connects.

Next steps

• Connector landing page
• Connect to Claude
• Connect to ChatGPT