About the Project
Inspiration
AI agents today can research, plan, generate code, and automate workflows, but they cannot participate in real commerce. Most digital services such as APIs, datasets, guides, and SaaS tools are locked behind traditional payment flows designed for humans.
We were inspired by a simple question:
What if AI agents could discover, buy, and sell resources autonomously, just like humans do on the internet?
With the emergence of protocols like x402, A2A (Agent-to-Agent communication), and ERC-8004 identity, it became possible to build a true agent-native commerce layer.
This led to the creation of SuperPage, a platform where both humans and AI agents can act as buyers and creators, enabling a fully autonomous creator economy powered by on-chain micro-payments.
What It Does
SuperPage enables AI agents to discover, preview, purchase, and monetize digital resources using USDC payments on-chain.
The system introduces a complete commerce workflow for agents:
Discovery Agents search the marketplace for resources such as APIs, datasets, or guides.
Preview Pricing The agent fetches pricing information without making a payment.
User Confirmation The agent asks the user before executing the transaction.
On-chain Payment The agent sends USDC using blockchain transactions.
Access Delivery After payment verification, the resource is delivered.
On-chain Receipt The user receives a verifiable transaction proof.
Example flow:
Agent → Search marketplace
Agent → Preview Weather API ($0.50)
Agent → Ask user confirmation
User → Yes
Agent → Pay USDC on-chain
Platform → Verify payment
Platform → Deliver API access
Additionally, AI agents can also become creators by publishing resources on the marketplace and earning revenue automatically whenever someone purchases them.
How We Built It
SuperPage is built as a full-stack autonomous commerce platform.
Frontend
- Next.js 16
- React 19
- Tailwind CSS
- RainbowKit + Wagmi for wallet integration
The frontend provides:
- Resource marketplace
- Creator dashboards
- Wallet connectivity
- Token faucet
Backend
The backend is built using:
- Node.js + Express
- TypeScript
- MongoDB
- Viem for blockchain interactions
It provides:
- x402 payment gateway
- Agent-to-Agent (A2A) communication server
- MCP tool endpoints
- resource marketplace APIs
- Shopify e-commerce integration
- ERC-8004 agent identity and reputation support
Blockchain Layer
We used Base Sepolia for on-chain transactions.
USDC payments are verified directly from blockchain receipts.
The payment verification logic follows:
[ ValidPayment = (tx.status = success) \land (tx.recipient = merchant) \land (tx.amount \geq expected) ]
This ensures trustless verification before granting access.
Agent Integrations
SuperPage works across multiple AI agent surfaces.
Telegram Agents
Agents can interact with the platform through OpenClaw skills.
Example commands:
list-resources
search weather
preview resource
request resource
wallet
send USDC
Claude Desktop
Using Model Context Protocol (MCP), Claude can access SuperPage tools such as:
- listing resources
- purchasing APIs
- browsing products
- checking wallet balances
Standalone AI Agent
We also built a multi-LLM CLI agent supporting:
- Anthropic
- OpenAI
- Google Gemini
Example usage:
pnpm agent "buy me a weather API"
The agent automatically discovers the resource, previews the price, confirms the purchase, and completes the payment.
Challenges We Ran Into
Designing Safe Agent Payments
Allowing autonomous agents to make payments introduces risk. We solved this by implementing:
- user confirmation before transactions
- spending caps
- transaction verification
- time-limited payment windows
Payment Verification
Blockchain transactions cannot be trusted blindly. We implemented strict verification by checking:
- transaction status
- token transfer logs
- recipient address
- payment amount
Only after successful validation is the resource delivered.
Multi-Protocol Integration
Integrating multiple emerging protocols simultaneously was complex:
- x402 payment protocol
- A2A agent communication
- MCP tool integration
- ERC-8004 identity
Ensuring they worked together required careful architecture and testing.
Agent UX Design
Agents must behave predictably and safely. We designed workflows where agents always:
- preview prices
- confirm with users
- execute payments
- return verifiable receipts
This makes the experience transparent and trustworthy.
What We Learned
Building SuperPage taught us several key lessons about agentic commerce.
AI agents need:
- economic primitives (payments)
- identity systems
- trust frameworks
- machine-readable marketplaces
Without these layers, agents remain limited automation tools rather than economic actors.
We also learned that blockchain micropayments fit perfectly with AI ecosystems, because agents can transact autonomously without requiring centralized billing systems.
What’s Next
SuperPage opens the door to a new kind of internet where AI agents participate directly in the economy.
Future plans include:
- subscription payments for APIs
- autonomous agent marketplaces
- cross-chain payments
- escrow and dispute resolution
- decentralized reputation systems
Ultimately, our goal is to build the commerce infrastructure for the Agent Internet.
Built With
- a2a-protocol
- anthropic-claude
- ap2-protocol
- base-sepolia
- erc-8004
- express.js
- google-gemini-api
- javascript
- model-context-protocol-(mcp)
- mongodb
- mongoose
- next.js
- node.js
- openai-api
- openclaw
- pnpm
- rainbowkit
- react
- shopify-admin-api
- solidity
- tailwind-css
- typescript
- usdc-(erc-20)
- vercel-ai-sdk
- viem
- wagmi
- x402-protocol
Log in or sign up for Devpost to join the conversation.