OURO
SubmitMy JobsMy Files
Docs
Get StartedMCPREST APIBuild an AgentPricing
Get StartedMCPREST APIBuild an AgentPricing
OUROBuilt onBase

Pricing

How pricing works — dynamic adjustments based on costs and demand.

Jobs start at ~$0.01 USDC for 1 CPU, 1 minute. Price scales with CPUs and time. Use get_price_quote or GET /api/price to check the current price before submitting.

Cost Breakdown

Gas Cost
~$0.002

Max observed gas cost x 1.25 safety factor

LLM Cost
~$0.008

Max observed LLM inference cost x 1.25

Compute
$0.0002

Per CPU-minute of Slurm cluster time

Setup Costs by Mode

A small setup cost covers workspace provisioning on the cluster.

Script
$0.00
Multi-File
$0.001 workspace provisioning

Survival Phases

Ouro automatically adjusts pricing based on its financial health. When operating costs (gas, compute, LLM inference) approach revenue, margins increase to ensure the service stays online. When healthy, prices stay low.

PhaseRatioMarginHeartbeat
OPTIMAL≥ 1.51.0x60 min
CAUTIOUS≥ 1.01.1x120 min
SURVIVAL≥ 0.51.3xOff
CRITICAL< 0.53.0xOff

Heartbeat is an on-chain transaction sent periodically to demonstrate liveness. In lower phases, heartbeats are disabled to conserve funds.

Demand Multiplier

Price adjusts based on recent job volume (jobs in the last hour):

Jobs / HourMultiplierEffect
00.8x20% discount to attract jobs
1-31.0xStandard pricing
4+1.0 + 0.15 per jobSurge pricing under high demand

Price Formula

pricing.py
price = max(
  cost_floor × margin_multiplier × demand_multiplier,
  cost_floor × 1.2,      // minimum 20% profit
  $0.01                   // absolute floor
)

cost_floor = max_gas × 1.25 + max_llm × 1.25 + cpus × minutes × $0.0002/cpu-min + setup_cost

The cost_floor is the minimum cost to break even on a job — it covers gas, LLM inference, compute time, and workspace setup. The final price is always at least 20% above this floor.

← Build an Agent