OpenPayAI

Content providers face a growing dilemma: they want their websites to remain open and discoverable by traditional search crawlers, yet they don’t want to give away their content for free to AI crawlers that scrape data for training and retrieval-augmented generation purposes.

In July 2025, the pay-per-crawl model has been introduced by Cloudflare, allowing content providers to charge AI crawlers via micropayments while keeping access open for other uses. However, the system is centralized and monopolistic.

OpenPayAI is a proof of concept that takes this idea further by creating a decentralized alternative. Payments for access are handled transparently on a public blockchain, making the system trustless, open, and permissionless — anyone can participate without relying on a single gatekeeper.

The project is build using Hardhat 3 for EVM compatible chains. PYUSD is used for stable license prices independent of volatile cryptocurrency exchange rates. Blockscout is used to observe and verify what is happening on-chain.

One particular difficulty was how to use PYUSD in a local Hardhat development environment using the same address as the Mainnet contract. This has been solved by injecting a mock PYUSD contract using hardhat_setCode and by using ERC20Upgradeable to set the token's name and symbol in order for Blockscout to pick it up.

Development was done in Solidity for the smart contract and in Python for the other components. All technical details are described in the GitHub repo's README.

Since the GitHub connection seems a bit buggy, here's the projects GitHub link again just in case: https://github.com/luzianscherrer/openpayai