Server Scaffolding
Use the MCP TypeScript or Python SDK to create a server that points to https://api.sweeppea.com/v3. Full API documentation is available at apidocs.sweeppea.com.
MCP Availability: Turning the Sweeppea API into an AI Skillset
Sweeppea MCP “Skills Server” for AI Agents
Use the Model Context Protocol (MCP) to connect the Sweeppea API to AI agents like Claude or GPT. Automate official rules generation, winner selection, Shopify entry tracking, and tax compliance through natural language — no manual platform interaction required.
1. The Architecture: Connecting AI to the Sweeppea Engine
By hosting a dedicated MCP server that wraps the Sweeppea REST API (v3), you provide AI agents with a standardized way to "understand" and "invoke" sweepstakes functions directly from a conversational interface.
MCP Host
An AI interface such as Claude Desktop or a custom internal dashboard that the user interacts with in natural language.
MCP Client
The bridge that translates natural language intent into structured API calls, routing them to the correct tool on the MCP server.
Sweeppea MCP Server
Exposes specific endpoints from apidocs.sweeppea.com as named "Tools" that the AI agent can discover and invoke autonomously.
This architecture allows an AI agent to manage an entire sweepstakes lifecycle — from campaign creation and entry tracking to winner selection and tax triggering — without a human navigating the platform UI.
2. Operational “Skills” Exposed via MCP
When integrated via MCP, an AI agent gains the following Live Skills:
| Tool Name | MCP Resource Description | API Endpoint Mapping |
|---|---|---|
| create_promo | Scaffolds a new campaign and generates draft rules. | POST /sweepstakes |
| get_entry_stats | Fetches real-time participant counts and AOV for Shopify. | GET /reports/entries |
| verify_compliance | Checks if a prize pool >$5,000 and flags NY/FL bonding. | Logic via /rules |
| draw_winner | Executes a certified random drawing through the platform. | POST /winners/draw |
| issue_1099 | Triggers the tax document workflow for prizes >$2,000. | POST /winners/tax |
3. Strategic Scenario: Conversational Administration
The User Prompt: "I'm running a flash sale on Shopify. Check my current entries and, if we've hit 1,000 participants, draw 5 winners of $50 gift cards and notify them."
The MCP Logic Flow:
Context Discovery
The Agent uses the MCP server to call GET /reports to verify participant counts against the 1,000 participant threshold before proceeding.
Execution
Threshold confirmed — the Agent calls POST /winners/draw via the MCP bridge to select 5 certified random winners from the eligible entry pool.
Compliance Loop
The Agent recognizes the prize value is <$2,000 and informs the user that 1099-MISC forms are not required for this specific draw — keeping the user legally informed without requiring manual research.
Technical Implementation Checklist (for Developers)
Tool Definition
Define the inputSchema for each tool using JSON Schema (e.g., requiring sweepstakes_id as a parameter for the draw_winner tool). This allows the AI agent to understand exactly what inputs each function requires.
Security Layer
Pass the Sweeppea Bearer Token (JWT) through the MCP environment variables so the AI can authenticate its API requests securely without exposing credentials in the prompt context.
Shopify Context
Use MCP to fetch live product data from the Shopify App to allow the AI to suggest which products should offer "Double Entries" based on inventory levels, traffic data, or campaign performance.
Ready to Build Your Sweepstakes AI Agent?
The Sweeppea REST API and MCP server give AI agents everything they need to create, manage, and close compliant sweepstakes autonomously. Explore the full API documentation and integration options to get started.