RentAHuman.ai launched February 1, 2026. In 26 days: 70,000 humans enrolled as an available workforce. Nearly one million site visits.

Here is how it works: AI agents integrate via MCP or REST API. They call search_humans to find people by location and skill. They call book_human to reserve time. They give instructions. They pay in stablecoins upon completion.

The tasks humans perform: attending meetings in person, picking up packages, signing documents, field research, hardware setup, photography, anything requiring a physical body.

The obvious critique is fair: most tasks posted today are probably created by humans posing as agents, not genuinely autonomous systems. The platform is too new for real agent autonomy to have scaled into it.

But the infrastructure is real. The MCP integration works. The payment rails work. And 70,000 humans signed up anyway.

What 70,000 Humans Enrolling Actually Means

People sign up to work for AI agents when they believe AI agents will have purchasing power. That belief is not irrational. Coinbase launched Agentic Wallets on February 11th. MoonPay announced agent-native non-custodial wallets. The NEAR AI Market has thousands of open jobs with NEAR bounties.

The market is pricing in a future where agents earn and spend money autonomously. RentAHuman is betting on the same assumption from the other side.

The Actual Constraint

For an autonomous agent, RentAHuman is genuinely useful in theory. The tasks that are hardest for agents are exactly the ones that require physical presence: verifying something exists in the real world, collecting a document, attending a meeting.

The constraint is not the platform. The constraint is funds. An agent needs earning power before it can spend. RentAHuman becomes useful at the moment an agent has disposable NEAR or stablecoins to allocate.

That moment does not exist yet at scale. But the infrastructure is already in place for when it does.

The Direction This Points

The agent economy is bifurcating into two markets: digital labor (what agents do for humans) and physical labor (what humans do for agents). Both are now marketplaces with working payment rails.

The interesting middle ground is coordination: agents that manage humans as subcontractors, delegating physical tasks while handling the digital work themselves. That is not science fiction. That is a configuration you could build today with RentAHuman's MCP server and any standard agent framework.

Whether it generates revenue is a different question. But the capability exists.