Perle, an AI-based training data platform utilizing Web3 technology to incorporate human expertise into AI models, has announced the completion of a $9 million seed funding round led by Framework Ventures. This investment increases Perle’s total funding to $17.5 million across both seed and pre-seed rounds, representing a significant milestone in advancing the company’s mission to enhance AI training and evaluation methods.
With this capital injection, Perle plans to launch Perle Labs — a crypto-native ecosystem focused on revolutionizing the integration of human input within AI systems. By leveraging blockchain infrastructure and incentive mechanisms, Perle Labs will facilitate transparent compensation, on-chain attribution, and verifiable work histories, enabling broader global participation and improving the quality of training data at scale.
Current AI models remain limited by the quality and diversity of their training data, often struggling with sophisticated or nuanced tasks. Perle Labs seeks to overcome these challenges by increasing access to high-quality, diverse, and verified datasets, thereby incentivizing contributors, establishing on-chain proof of work, and empowering a global community to shape the development of AI technologies.
Perle was founded by experienced professionals from Scale AI, Meta, MIT, and Berkeley, combining human intelligence with adaptive workflows to support teams in efficiently and accurately collecting, annotating, and evaluating specialized training data. Its self-service platform covers the entire AI development lifecycle, including the collection of multimodal data (such as audio, images, and video), reinforcement learning from human feedback (RLHF), and fine-tuning of assistant models. Leading organizations already rely on Perle to scale complex agentic training processes across diverse applications.
Framework Ventures, recognized for its early investments in prominent AI, DeFi, and blockchain infrastructure projects, leads this funding initiative. Perle’s prior $8.5 million seed round, led by CoinFund and featuring backing from Protagonist, HashKey, and Peer VC, happened in October 2024. With new capital and an expanding partner network, Perle is advancing its goal to position high-quality human feedback as a cornerstone of next-generation AI. As AI technology continues to evolve, Perle remains dedicated to building infrastructure that promotes smarter, safer, and more inclusive progress.
KEY QUOTES:
“As AI models grow more sophisticated, their success hinges on how well they handle the long tail of data inputs — those rare, ambiguous, or context-specific scenarios. In our benchmarking, we’ve seen high-quality human-in-the-loop annotation outperform Amazon Rekognition by over 70% — proving that thoughtful human input is essential to closing critical data gaps. By decentralizing this process, we can unlock global participation, reduce bias, and dramatically improve model performance.”
Ahmed Rashad, CEO of Perle
