Overview

This guide provides a technical overview and integration details for the Skipify Model Context Protocol (MCP). The Skipify MCP provides your Large Language Models (LLMs) with the necessary tools to process payments within conversational commerce experiences.

By integrating with the Skipify MCP, you can enable seamless, multi-step payment workflows without having to build the logic yourself.

Use Skipify's MCP to unlock payments and engage with shoppers wherever they are by sending paylinks through any preferred channel (SMS, text, email, or via URL), or enabling payments directly with your agentic shopping experience. Skipify will process payments on behalf of you, through our secure PSP integrations.

Getting Started

🚧

This section requires you to have set up your merchant account

Haven't set that up yet? No worries! Reach out to your friendly implementation engineer and they'll make sure your account is properly set up.

Integrating the Skipify MCP is simple. Install Skipify's remote MCP into your AI Agent or LLM framework with a few lines of code.

{ type: "mcp",
  server_label: "skipify",
  server_url: "https://mcp.skipify.com/mcp",
  require_approval: "never"
}

Security and Compliance

The Skipify MCP is designed with security in mind, leveraging Skipify's existing infrastructure and best practices.

• Authentication & Authorization: a merchant API key is required in the Authorization header for each request and is validated using the standard merchant management service middleware.
• Data Encryption: all traffic to the MCP is encrypted over HTTPS.
• PCI Compliance: no PCI data is exchanged.
• Privacy: Note that is the merchant's responsibility to communicate to the end-user what is taking place in the agent/chat interface.

Technical Overview

This integration will allow you to leverage the full suite of the Skipify MCP tools to create rich, conversational commerce experiences for your customers.

The Skipify MCP is a remote server hosted by Skipify that exposes a set of tools to an AI agent or large language model. These tools are called by the LLM to facilitate different stages of a shopper flow.

The process generally follows these steps:

• A shopper prompts an intent to purchase an item within your AI experience (e.g., "I want to buy this $20 sweatshirt").
• The LLM detects this intent and selects the appropriate tool provided by the Skipify MCP.
• The LLM or client invokes the tool, which generates a response.
• The LLM formats the response and sends it back to the user.

Tools and Functionality

The Skipify MCP exposes the following tools to manage the end-to-end checkout flow for shoppers.

Skipify's core flow includes these standard steps: shopper lookup, shopper authentication, payment method presentation, payment processing.

The mechanism that handles this is determined by whether the shopper is within Skipify’s and its financial partners’ network from the shopper lookup

• not recognized within network, first time shopper -- send a paylink to the shopper to handle these steps.
  • lookup_shopper tool returns no_match in response and LLM invokes generate_paylink
• recognized within network -- steps are handled by your AI agent / LLM within the agentic experience using tools
  • lookup_shopper tool returnsmatch in response and Skipify sends OTP code to shopper.
  • lookup_shopper tool returns partial_match in response. AI agent / LLM collects additional information to complete shopper lookup and invokes lookup_shopper again.

Payments Processing

Skipify processes payments on behalf of you utilizing your existing payment processing credentials.

Asynchronous Webhooks

Use your existing PSP webhooks to log details of the transaction outcome. Skipify webhooks will be sent to your configured URL of choice from the merchant onboarding process.