MCP Server for Veryfi Document Processing

MCP Server for Veryfi Document Processing

Overview

A Model Context Protocol (MCP) server implementation that integrates with Veryfi for data extraction from documents

Setup

Get your Veryfi username, client_id, and api key

Log into https://app.veryfi.com, then navigate to https://app.veryfi.com/api/settings/keys/

With Claude Desktop

Modify claude_desktop_config.json to include this:

Do not set environment variables here. It will cause confusion if the enviroment variables used by the MCP server differ from those used by your other code.

Instead, setting them in your execution environmont or using a .env file.

{
    "mcpServers": {
        "veryfi": {
            "command": "uv",
            "args": [
                "--directory",
                "path/to/mcp-server",
                "run",
                "src/server.py"
            ],
            "env": {}
        }
    }
}

Development Setup

Local Development

1. Clone the repository

2. Create a .env file with the following variables:

   VERYFI_CLIENT_ID="veryfi-client-id-goes-here_client_id"
   VERYFI_USERNAME="veryfi-username-goes-here"
   VERYFI_API_KEY="veryfi-api-key-goes-here"
   

Alternatively, you can set your environment variables in your execution environment. For example, on a Mac, append this to your ~/.zprofile

export VERYFI_CLIENT_ID="veryfi-client-id-goes-here_client_id"
export VERYFI_USERNAME="veryfi-username-goes-here"
export VERYFI_API_KEY="veryfi-api-key-goes-here"

3. Install dependencies:

   pip install uv
   uv sync
   

4. Run the server locally:

   uv run mcp dev src/server.py
   

Testing

Run the test suite with:

uv run pytest