The problem ZKp Plagiarism detector solves

Plagiarism is a critical issue in academia, content creation, and intellectual property protection. Traditional plagiarism detection methods expose document content to centralized services, raising concerns about data privacy, trust, and ownership verification. This project addresses these challenges by:

Ensuring Privacy – The zk-proof mechanism allows verification without revealing document content.

Preventing Fraud – Immutable blockchain records ensure that originality claims cannot be falsified.

Eliminating Centralized Trust – No need to rely on third-party plagiarism checkers.

Enabling Proof of Authorship – Writers and researchers can prove their work’s originality on-chain.

Providing Verifiable Certificates – NFT-based certification enables transparent validation of original work.

Challenges I ran into

One of the major challenges I faced was the lack of decentralized ML models for embedding generation and similarity comparison. Since no fully on-chain solution exists for this task, I had to implement a Python server to handle these computations. This introduced a degree of centralization, which conflicts with the core principles of decentralization.

Another challenge was setting up a zkVM as a server to expose an endpoint for accepting similarity scores. Since this required working with Rust-based zkVM implementations, integrating it with the rest of the system and ensuring smooth communication between components added complexity.