Skip to main content
The Vero MCP Server is a remote Model Context Protocol server that lets AI assistants interact with your bank accounts, transactions, and receipts through Vero’s platform. Instead of building custom integrations, any MCP-compatible client — Claude Desktop, Claude Code, Cursor, or others — can connect to Vero’s hosted server and immediately access a full suite of financial tools.

Quick setup

Connect your AI assistant to Vero in under 2 minutes.

Tool reference

Browse all available tools with parameters and examples.

Authentication

Understand the OAuth 2.0 flow that secures every request.

​
What you can do

Vero attaches itemized line items to the actual transactions in your bank accounts, turning every purchase into a detailed, queryable record. Once connected, your AI assistant can:
  • Query purchases at the line-item level — Ask “how much did I spend on coffee last month?” or “what did I buy at Target in March?” and get answers from the actual items on each receipt, not just merchant totals
  • Generate expense reports — Group, categorize, and export itemized spend across transactions for reimbursements or client billing
  • Prepare tax-ready records — Surface deductible purchases with itemized proof attached to the transaction they belong to
  • Build spending awareness — Break down where your money goes by category, merchant, or specific product — across all connected accounts
  • Attach receipts to transactions — Upload images via OCR, or connect Gmail to auto-extract receipts and match them to the right transaction

​
Architecture

The Vero MCP Server sits between your AI assistant and the Vero backend, handling authentication and translating MCP tool calls into API requests.

​
How it works

1

Client connects

Your MCP client sends a request to the /mcp endpoint. If no Bearer token is present, the server responds with a 401 and a discovery URL, triggering the OAuth flow.
2

User authenticates

The client follows the OAuth 2.0 authorization code flow through Auth0. The user logs in (or signs up) and grants access. The client receives a Bearer token.
3

Tools become available

With a valid token, the MCP client discovers all available tools and their schemas. The AI assistant can now call any tool on behalf of the user.
4

Requests flow through

Each tool call is authenticated, forwarded to the Vero API with the user’s token, and the result is returned as structured JSON to the AI assistant.

​
Technical details

​
Tool categories

The server exposes tools across 5 categories: See the complete tool reference for details on every tool and its parameters.