AgentPay

Inspiration

The gig economy is evolving. We believe the next workforce won't just be human freelancers found on Upwork or Fiverr; it will be Autonomous AI Agents. However, current payment rails are too slow and riddled with fees for this high-speed, micro-transactional future.

We asked ourselves: "What if you could instantly hire a specialized AI expert—like a Smart Contract Auditor or a Financial Analyst—and pay them seamlessly on-chain?"

This question birthed AgentPay: A marketplace for Programmable Intelligence, powered by the speed of MNEE and the reasoning capabilities of Gemini 2.5.

What it does

AgentPay is a decentralized platform where users can browse and "hire" specialized AI agents for specific tasks.

• Ledger: A financial analyst for market trends.
• Syntax: A smart contract auditor for vulnerability checks.
• Muse: A creative copywriter for Web3 marketing.

Users pay a one-time fee in MNEE to unlock the agent's services. The payment is settled instantly on-chain, and the agent (powered by Gemini 2.5 Flash) immediately executes the task, delivering professional-grade results in seconds.

How we built it

We built AgentPay with a focus on speed and low latency to match the "instant" nature of AI.

• AI Engine: We utilized Google Gemini 2.5 Flash for the agents' brains. Its low latency and high reasoning capabilities were essential for creating agents that feel "alive" and responsive.
• Payments: We integrated MNEE Stablecoin as the exclusive currency. Its speed allows for a friction-free "Pay & Use" experience that traditional crypto or fiat rails can't match.
• Frontend: Built with React and TailwindCSS for a modern, dark-mode aesthetic that fits the Web3 vibe.
• Backend: A lightweight serverless architecture that handles the handshake between the on-chain payment confirmation and the Gemini API request.

Challenges we ran into

One of the main challenges was orchestrating the "Payment-to-Action" flow. We had to ensure that the AI agent only triggers after the MNEE transaction is fully confirmed on-chain. Balancing the wait time for blockchain confirmation with the user's expectation of an "instant" AI response required fine-tuning our listeners.

Accomplishments that we're proud of

We are proud to have built a working "Service-as-a-Software" model. Successfully connecting a live crypto wallet payment (MNEE) to trigger a specific, prompt-engineered AI personality (Gemini) feels like stepping into the future of the internet.

What we learned

We learned that Gemini 2.5 Flash is incredibly adept at role-playing specific professional personas (like an Auditor vs. a Copywriter) when given the right system prompts. We also realized that stablecoins like MNEE are the only viable path for the future of Machine-to-Machine (M2M) commerce.

What's next for AgentPay

• Agent-to-Agent Payments: Allowing our "Muse" agent to hire the "Syntax" agent to check its own work.
• Subscription Models: Recurring MNEE payments for long-term agent retention.
• Mainnet Launch: Moving from the hackathon demo to a live public marketplace.

Built With

• ai
• blockchain
• gemini
• mnee
• node.js
• react
• stablecoin
• tailwindcss