Pay for outcomes. Not for seats. Not for tokens.

AgentSpot is based on DO credits: one credit per outcome. Not per user who might log in. Not per million tokens the model happened to consume. Tokens still power AgentSpot under the hood — we just don't make them your problem.

How Is AgentSpot Priced Differently?

DO credits agentspot-logo
Per user (others)Per token (others)
What you are paying forWork completedA user with a loginModel consumption
Cost of adding your whole team$0Grows with every seatN/A
Cost of a complex taskSame as a short oneUnpredictableGoes up, unpredictably
Cost when a run fails$0You still payYou still pay
What you forecast againstOutcomes you already planHeadcount you do not controlToken math you cannot predict

One workflow — 15 tasks

AgentSpot
1DO CREDIT
Other agentic tools
1000sOF TOKENS

What Uses A DO Credit

EventCost
New agent conversation is started1 DO credit
A workflow run completes1 DO credit
An application is loaded1 DO credit

What Never Uses A DO Credit

Building
Building

Create, test and iterate on agents, workflows and apps, free on both Free and Fleet.

Failures
Failures

A workflow that does not complete, an app that does not load: no charge. You pay only when it works.

People
People

Unlimited users. No builder seats, no viewer seats, no per-user subscriptions.

What That Looks Like In Practice

Scenario: a 150-person ops team running 250 workflows per week (~1,000 per month)

AgentSpotPer-seat toolPer-token tool
How you're billedBy workflow run: 250 DO credits per weekBy headcount: 150 seats × $20/monthBy consumption: ~1,000 workflows × $3–$5 each
Monthly cost$1,650$3,000$3,000–$5,000