ArcPay: Engineering Autonomous Financial Workflows with Gemini & Circle

ArcPay: Engineering Autonomous Financial Workflows with Gemini & Circle


AI Hackathon Project
AI Agents FinTech Web3

Today, AI agents are “Financial Orphans.” They can plan and reason, but they struggle to move value autonomously due to complex cash management and unreadable blockchain data. During the Agentic Commerce on Arc Global Hackathon (organized by Lablab.ai in collaboration with Google DeepMind & Circle), I developed ArcPay—a financial nervous system for autonomous agents.

The Vision: Financial Autonomy

The goal of ArcPay was to bridge the gap between Large Language Models (LLMs) and programmable finance. In a world of “Agentic Commerce,” we need systems where an AI can not only find a product but also procure it securely without a human having to copy-paste wallet addresses or sign every transaction manually.

Technical Architecture

I built ArcPay using a modular stack designed for speed and transparency:

  1. Circle Developer Wallets: Leveraged Circle’s infrastructure to create programmable, secure wallets that an AI agent can interact with via API.
  2. Google Gemini 1.5 Flash: Used to solve the “Transparency Gap.” Gemini interprets raw transaction hashes and translates them into human-readable summaries, providing a clear audit trail for the user.
  3. The Arc Network: Integrated the Arc protocol to ensure high-speed finality for agent-to-agent transactions.

Key Technical Achievements

  • Agentic Intent Interpretation: Built a backend that allows the AI to understand financial intent—translating a user’s prompt like “Send 5 USDC for the monthly subscription” into a verified transaction.
  • Automated Financial Audits: Integrated a feature that allows users to click “Explain with AI” to get an instant, jargon-free breakdown of any blockchain movement.
  • Scalable Prototyping: Developed a modular Python-based backend that demonstrates how decentralized commerce can be simplified for the average user.

Reflection: The Value of Global Collaboration

While the project was built in the high-pressure environment of a global hackathon, the true value was in the technical iteration. Navigating the documentation for Google’s Gemini and Circle’s SDKs simultaneously provided immense insight into the future of the Autonomous Economy. ArcPay serves as a blueprint for how we can build trust in systems where AI makes the decisions.


#AgenticCommerce #GoogleGemini #Circle #Blockchain #AI #Fintech #Web3 #Lablabai

© 2026 Noor Ahmed
Software Engineer • Open Source Contributor • Social Worker • Content Creator