vivid

Vivid MCP provides a remote Model Context Protocol (MCP) endpoint you can connect to from any MCP-compatible client (Claude, ChatGPT, Cursor, etc.). The server acts as a bridge to Vivid's money-related capabilities, enabling you to instruct an AI to open a Vivid Business account by supplying company details in chat. Once connected, the MCP client streams requests to the endpoint, which responds with structured actions and results that the AI can interpret to complete onboarding steps. Typical use cases include uploading or describing your company documents, triggering account creation, and guiding the user through identity verification and pre-fill workflows. The README indicates that the service exposes a single, centralized endpoint at https://api.prime.vivid.money/mcp, which clients can integrate with without an API key.

To use it, configure your MCP client to target the vivid endpoint, then start the local/mapped server (as shown in your environment). The client will send user intents (e.g., open a business account, fetch account status) via the MCP protocol, and the vivid-mcp backend will respond with the appropriate actions and data needed to complete the interaction.

Prerequisites:

• Node.js installed (recommended for the provided example configuration)
• Access to the Vivid MCP endpoint (https://api.prime.vivid.money/mcp)

1. Prepare the environment

• Create a project directory and navigate into it
• Ensure you have a server script that can connect to the MCP endpoint. The following example assumes a Node.js-based server at path/to/server.js.

2. Install dependencies (if applicable)

• If your server requires dependencies, initialize a package.json and install them, e.g.:

npm init -y
npm install axios  # or any MCP-compatible library you plan to use

3. Configure the MCP server entry

• Create or edit your server script (path/to/server.js) to connect to the MCP endpoint:

// Example outline (pseudo-code)
const mcpEndpoint = process.env.MCP_ENDPOINT || 'https://api.prime.vivid.money/mcp';
// initialize MCP client and connect

4. Run the MCP server locally

node path/to/server.js

5. Verify connectivity

• Ensure the server prints a ready/status message and that a test client can connect to the endpoint.

Note: If you are using a different runtime (Python, uvx, or a docker image), adapt the commands accordingly to run a script that connects to https://api.prime.vivid.money/mcp and exposes the MCP interface to your client.

Tips and caveats:

• The Vivid MCP endpoint does not require an API key, but you should monitor for rate limits or policy changes in the future.
• If you deploy behind a firewall or in a restricted environment, ensure outbound HTTPS access to api.prime.vivid.money/mcp is allowed.
• Keep your MCP client updated to support the latest MCP protocol features and any endpoint changes.
• Since the README describes various integration options (Claude, ChatGPT, Cursor, Replit, etc.), ensure you configure your client according to the respective platform’s instructions for connecting to an external MCP endpoint.
• If you encounter connectivity issues, verify that MCP_ENDPOINT is correctly set and that TLS certificates are trusted in your runtime.