A commerce-centric demo featuring Agent Development Kit with Paypal Agent Toolkit via MCPs
@gkcng
关于 A commerce-centric demo featuring Agent Development Kit with Paypal Agent Toolkit via MCPs
The commerce-centric demo shown live at Paypal Dev Days 2025 on 4/29. Featuring Paypal Agent Toolkit & MCP server with Google's Agent Development Kit (ADK)
基本信息
配置
工具
未检测到工具
工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。
概览
What is A commerce-centric demo featuring Agent Development Kit with Paypal Agent Toolkit via MCPs?
This demo showcases a conversational pre-sales engagement built with PayPal's Agent Toolkit (via MCP) and Google's Agent Development Kit (ADK). It is intended for developers exploring agent-based commerce workflows using the PayPal sandbox environment.
How to use A commerce-centric demo featuring Agent Development Kit with Paypal Agent Toolkit via MCPs?
Install prerequisites: Python 3, Google ADK 0.3.0+, Node.js with npx, and the required Python packages. Copy env.example to .env, fill in your Google API Key (or GCP Project ID/Location) and PayPal sandbox credentials (PAYPAL_CLIENT_ID, PAYPAL_SECRET). Modify the user_profile and orders dictionaries in paypal_mcp_agent/util.py with real sandbox order data. Run adk web in the pp_mcp_adk directory, then open http://localhost:8000/dev-ui?app=paypal_mcp_agent in a browser.
Key features of A commerce-centric demo featuring Agent Development Kit with Paypal Agent Toolkit via MCPs
- Demonstrates pre-sales conversational agent flow with PayPal APIs.
- Integrates Google ADK with PayPal Agent Toolkit via MCP.
- Covers order creation, tracking, shipment, and subscriptions.
- Includes product search and image-based recommendation.
- Uses PayPal sandbox for testing without real transactions.
- Supports A2A for adding remote external agents.
Use cases of A commerce-centric demo featuring Agent Development Kit with Paypal Agent Toolkit via MCPs
- Build a conversational agent that handles order inquiries and tracking.
- Create a pre-sales assistant that recommends and creates orders from images.
- Test PayPal commerce APIs (orders, subscriptions, disputes) in a sandbox.
- Extend the agent with remote agents via A2A for specialized tasks.
- Learn how to integrate Agent Development Kit with PayPal MCP tools.
FAQ from A commerce-centric demo featuring Agent Development Kit with Paypal Agent Toolkit via MCPs
What are the runtime requirements?
Python 3 with Google ADK 0.3.0+ (google-adk==0.3.0, google-cloud-aiplatform==1.90.0) and Node.js with npx for the MCP client. A PayPal sandbox account is required.
How do I set up the PayPal sandbox?
Register at developer.paypal.com to create apps. Each app provides a Client ID, Secret Key, and a sandbox account with test credit cards, addresses, and bank details for testing transactions.
Which PayPal APIs are used in this demo?
The demo tests and uses these APIs: orders.create, orders.get, orders.capture, shipment.create, shipment.get, subscriptions.create, subscriptions.show, subscriptions.cancel.
What are important API interdependencies?
Order capture succeeds only after the consumer completes payment via the returned URL. Subscriptions require a previously created product and plan. Disputes must be created by the human in the sandbox before a merchant can accept them. list_transactions has a delay.
Where does the agent’s data live?
The demo loads a hardcoded user_profile and orders dictionary from paypal_mcp_agent/util.py. Order data must correspond to actual records in the PayPal sandbox. No persistent database is used.
AI 与智能体 分类下的更多 MCP 服务器
LinkedIn MCP Server
stickerdanielOpen-source MCP server for LinkedIn. Give Claude and any MCP-compatible AI agent access to profiles, companies, jobs, and messages.
Gemini MCP Server
aliargunMCP server implementation for Google's Gemini API
Intervals.icu MCP Server
mvilanovaModel Context Protocol (MCP) server for connecting Claude and ChatGPT with the Intervals.icu API.
Solon Ai
opensolonJava AI application development framework (supports LLM-tool,skill; RAG; MCP; Agent-ReAct,Team-Agent). Compatible with java8 ~ java25. It can also be embedded in SpringBoot, jFinal, Vert.x, Quarkus, and other frameworks.
meGPT - upload an author's content into an LLM
adriancoCode to process many kinds of content by an author into an MCP server
评论