Human API

Human API has launched a platform that enables artificial intelligence agents to directly assign and manage tasks performed by humans. This system is designed to address the 'last mile' problem in AI, where autonomous agents are unable to complete real-world tasks that require physical interaction or nuanced human judgment. The company was developed by Eclipse, the team behind an Ethereum Layer 2 built on the Solana Virtual Machine.

The platform functions as an agent-native coordination and execution layer. AI systems can programmatically generate tasks, which are then routed to human contributors who have created accounts on the platform. These tasks can range from capturing high-quality audio that includes specific accents and linguistic nuances to actions requiring physical presence, such as logistics and deliveries. After a human completes a task, the submission goes through a review process, and upon approval, the contributor is paid via Stripe Connect. This creates a marketplace where AI companies can access scalable human input, and individuals can monetize skills that are difficult to automate.

As of February 2026, Human API has exited stealth and raised $65 million in funding from investors including Placeholder, Hack, Polychain, DBA, and Delphi Ventures. The company's initial business model focuses on licensing high-fidelity audio datasets to AI labs. In its stealth phase, Human API contributed to the creation of a studio-quality audio dataset for a frontier AI lab. Future plans include expanding into other data types like computer-usage data and tasks that require real-world execution.

Historically, the name Human API was associated with a health data platform founded in 2013 by Andrei Pop and Ola Wiberg. That company created a consumer-controlled platform for aggregating and sharing health data from thousands of sources, including hospitals, labs, and wearables. It was acquired by LexisNexis Risk Solutions in April 2023. The new iteration of Human API, launched in February 2026, is a distinct entity with a new founder, Sydney Huang, and a completely different focus on bridging AI agents with human labor.

Keywords: AI agent coordination, human-in-the-loop, agent-native system, real-world task execution, AI task marketplace, human-generated data, audio data licensing, AI logistics, data-as-a-service, human computation, decentralized workforce, Eclipse, Sydney Huang, task outsourcing, programmatic workflows, verified contributors, on-demand human tasks, AI-human collaboration, physical observability, multilingual voice models, autonomous systems support