The problem DeMol solves

Traditional drug discovery is slow, expensive, and centralized—often requiring years of lab work and millions in funding before a molecule even reaches preclinical validation. Promising discoveries made by independent researchers frequently go unfunded or unnoticed due to lack of access to institutional support.

DeMol solves this by:

• Enabling AI-powered molecule generation and evaluation in minutes rather than months.
• Providing a platform for scientists and researchers to submit, score, and showcase promising compounds without needing a big pharma pipeline.
• Tokenizing high-quality molecules as on-chain IP, making them tradable, fundable, and verifiable from day one.
• Letting a decentralized community (DAO) vote on which molecules deserve funding—creating a global, democratized R&D engine.
• Streamlining the IP lifecycle, from generation to validation and funding, using automation, verifiable AI, and smart contracts.

What people can use it for:

• Researchers can validate and monetize their discoveries.
• DAOs and investors can fund molecules and earn royalties from future licensing.
• Developers can contribute AI agents, tools, or data pipelines.

Challenges we ran into

• ELiZA Inference Mapping: Customizing agent actions to be recognized and executed correctly within ELiZA’s framework.
• Lightweight Deployment: Running ELiZA agents on small servers (e.g., AWS micro instances) caused performance bottlenecks.
• Toxicity Model Balance: Managing performance across multiple toxicity endpoints with limited data.