Clarylisk

Demo

About

The problem it solves Clarylisk solves a critical problem in the creator economy: the abuse of donation platforms to spread harmful messages, particularly online gambling promotions. On traditional platforms, donations often come with custom messages that appear publicly during livestreams. Unfortunately, this system is vulnerable to misuse—bad actors exploit it by attaching gambling-related promotions, damaging the creator’s reputation and putting their audience at risk. Existing donation platforms don’t give creators control to filter or reject such messages before they're displayed. Once a message appears on screen, the harm is already done. Clarylisk introduces a Web3-based donation platform that integrates blockchain transparency with AI-powered moderation. Every donation is analyzed by an AI model trained to detect gambling-related content. If a message is flagged, it enters a pending state, giving creators the power to approve or burn the donation before it's finalized. This mechanism restores control to the creator, discourages harmful actors, and offers a transparent dashboard to prove that the creator maintains ethical standards. Through Clarylisk, creators gain safety, accountability, and trust — setting a new standard for donation integrity in the Web3 era.

How it's made

Challenges we ran into One of the biggest challenges we faced while building Clarylisk was integrating the AI moderation system seamlessly into the decentralized donation workflow. Since every donation involves an on-chain transaction, we needed a way to scan messages in real time without compromising the speed or trustless nature of the platform. Initially, we tried embedding the AI model directly into the frontend, but it significantly affected performance and caused inconsistencies due to client-side processing limitations. To overcome this, we shifted the architecture by offloading AI inference to a backend API that runs a lightweight and optimized version of the model. This API acts as a bridge — donations are first submitted off-chain to our backend, scanned, and only then passed to the blockchain with a flag indicating whether it’s clean or suspicious. This allowed us to maintain both efficiency and creator control, while keeping the dApp responsive. Another hurdle was building a burn mechanism on-chain that could reject donations post-submission. Designing a contract that allowed fund rejection while preserving blockchain integrity required several iterations. We eventually implemented a system where flagged donations are held temporarily in a contract and only transferred after creator approval. If rejected, the funds are permanently removed from circulation — reinforcing the platform's stance on ethical donations. These technical and ethical hurdles helped us sharpen the product’s direction and strengthen its value proposition as a secure, AI-integrated donation platform.