Our journey with OnchainKit started smoothly—integrating it with Smart Wallet for wallet connections and signing was seamless, which we appreciated as part of the Base ecosystem. However, we encountered significant challenges when trying to go beyond these basics, particularly with testing on Base Sepolia. Despite our best efforts, we couldn’t find any documentation or guidance on configuring OnchainKit for Base Sepolia. This lack of support for the testnet environment was a major roadblock, as testing with real USDC on Base mainnet was cost-prohibitive for our team.

To keep development moving, we devised a workaround: deploying a mock USDC ERC-20 token on Base mainnet. As admins of this token, we could mint as needed, simulating funding flows without financial risk. While this allowed us to continue building, it meant we couldn’t fully leverage OnchainKit’s funding and buying features, limiting the scope of our prototype. This detour underscored the need for better testnet support in OnchainKit to help developers iterate without mainnet costs.

Another challenge was our lack of smart contract expertise. With no dedicated contract developer on the team, we leaned heavily on AI tools like Ohara and Cursor. Ohara helped us generate an initial MVP contract, while Cursor refined the code and guided us through deployment and verification on Base mainnet. These tools were invaluable, but we still face the task of implementing complex features like milestone-based funding and voting logic—key to our vision but not yet realized in the contract.

These challenges pushed us to be resourceful and creative, reinforcing the value of AI-driven development in overcoming barriers. They also highlighted areas where Base’s tools, like OnchainKit, could evolve to better support builders, especially in testnet environments.