Image Credit: Shutterstock
Circle has launched Arc Studio, an AI coding agent that turns natural-language prompts into full-stack on-chain applications, smart contracts and software agents.
The tool follows Arc’s public mainnet launch and integrates with Circle infrastructure including CCTP, Gateway, Wallets, and Contracts. Developers can generate frontend interfaces, backend logic and Solidity contracts from a written description.
Arc Studio Generates Full-Stack Apps From Natural-Language Prompts
Developers can describe an application in plain English and have Studio assemble a working prototype around the request. Generated projects can include wallets, payments, swaps, bridges, lending and programmable settlement.
Studio can also inspect existing smart contracts from an address and generate interfaces around them. Its tracing tools record RPC calls, transactions, decoded events and off-chain requests while developers test applications.
The tool works through a web interface and can also operate as a subagent inside Claude Code, Codex, and Cursor. Generated code can be exported into a developer’s own repository for review and further development.
Developers Can Test Across Nine EVM Networks
Arc Studio supports testing across Arc, Ethereum, Base, Arbitrum, Avalanche, Monad, OP Mainnet, Polygon and Unichain.
The agent can generate Solidity contracts, perform an initial code review and deploy them to Arc Testnet or eight other EVM testnets. It also integrates with protocols including Aave, Morpho and Uniswap.
Production deployment remains separate. Arc Studio does not deploy, sign, fund or submit mainnet transactions on behalf of users. Developers must use their own credentials and infrastructure when moving generated applications into production.
Circle Targets USDC Payments and Agent-Based Applications
Circle is positioning Studio around financial applications involving USDC, including payments, swaps, treasury workflows and software-agent transactions.
The launch follows Arc’s Sept. 16 public mainnet debut, which brought more than 100 applications and more than 100 institutional and ecosystem builders onto the network.
Circle warns that AI-generated code may be inaccurate, incomplete or contain bugs and security vulnerabilities. Developers remain responsible for independently reviewing, testing and validating generated applications before production use.
