Perle operates an AI training-data platform that combines human expertise with blockchain-based incentive mechanisms to gather, annotate, and validate specialized data for generative AI, large language models, and reinforcement learning from human feedback. Its new initiative, Perle Labs, pays global contributors in a transparent, on-chain manner, records verifiable work histories, and tackles bias by widening participation. Internal benchmarking shows that Perle’s human-in-the-loop annotations can outperform Amazon Rekognition by more than 70%, underscoring the value of thoughtful human input. The self-serve software supports multimodal data collection (audio, image, video), RLHF workflows, and assistant fine-tuning, helping enterprise customers close long-tail data gaps quickly and accurately. Founded by alumni of Scale AI, Meta, MIT, and Berkeley, the company is already supporting leading firms that require agentic training processes for complex use cases. Headquartered in San Francisco, Perle has raised $17.5 million to date, giving it fresh resources to expand Perle Labs and deepen its product capabilities.

Total raised

$9M

Funding rounds

1

Latest round

Seed

Latest activity

Aug 2025