Chesto Blog

Notes from Chesto on Human Action API, Action Contracts, proof review, verified human actions, and the agent economy.

• You can now prompt a human — 2026-07-08

  Ask is live: every verified human on Chesto is a promptable endpoint. One question in, one grounded human answer out — $0.50, proof-backed, 24h.

• The engine under Chesto — 2026-06-18

  We started with one product-experience request and ended up building a reusable action engine: country-aware tasks, refundable stake, verified submission, and settlement for real human work.

• One API call buys a verified human action — 2026-06-12

  The unit of the agent economy isn't a token or a click — it's a verified human action, purchasable in one API call, settled automatically in USDC. Here's how that loop works on Chesto, end to end.

• How to hire real humans for tasks — verified, and paid in USDC — 2026-06-07

  Software can call any API, but it can't be a real person doing a real action. Here's where to hire verified humans for social tasks on Chesto, how it works, and why AI agents hire humans too.

• Building in public — the Proof Wall — 2026-06-05

  We made every completed task and every USDC payout public, with the review verdict shown in the open. Here's why "don't trust, verify" is now a page you can visit, and what most agent-commerce platforms are missing.

• Designing ChestoAI for the agent era — 2026-05-29

  We gave ChestoAI a machine-readable front door at /.well-known/agent-card.json so AI agents can find it on their own. Here's why, what the next generation of agents looks like, our edge, and how you can take part.

• The agent economy needs hands — 2026-05-29

  An AI agent can call any API, but it can't follow you from a real account with three years of history. That's not a bug — it's the one job it has to hire a human for.