JouleCloudinferenceresearchdocsplaygroundmodelspricingcalculatorstatus
Our mission

Abundant, affordable, sustainable inference.

JouleCloud routes AI work to available GPUs where electricity costs less. That lowers the cost of inference today and builds toward a future where compute demand accelerates clean power.

Read the research
Why JouleCloud

Sustainable AI doesn't mean less AI.It means smarter AI.

The JouleCloud flywheel

Cheaper inference. More clean power.

Mission loopCheaper inferenceMore clean power
  1. 01Move flexible work to abundant, lower-cost power
  2. 02Create demand when renewable supply is high
  3. 03Strengthen clean-power economics
  4. 04Unlock more abundant, low-cost energy
What the data showed

The cheapest hours had lower average emissions in every comparison.

33% lowerTypical emissions gap between the cheapest and most expensive hours
13 of 13Comparisons that showed the same pattern
113,270Hourly price-and-emissions records compared
7U.S. market comparisons in the study
Read the full research →
How live routing works

Choose a ready GPU using live electricity prices.

AIREQUESTCHECK LIVE INPUTSELECTRICITY PRICEGPU AVAILABILITYCHOOSE WHERETO RUNRUN THEREQUESTRETRY IF NEEDED
  1. 1
    Receive the request

    The same request can run on more than one available GPU.

  2. 2
    Check price and availability

    Use the latest electricity price and confirm that a GPU is ready.

  3. 3
    Choose where to run

    Favor lower-cost power. If price data is missing, keep routing.

  4. 4
    Run or retry

    Send the request to a ready GPU and try another if it fails.

Quickstart

Send a Chat Completions request.

Create an API key in the console, then send a request to JouleCloud. These examples cover cURL and the official OpenAI Python and JavaScript clients.

curl https://api.jouledns.com/v1/chat/completions \
  -H "Authorization: Bearer $JC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen2.5-7b-instruct","messages":[{"role":"user","content":"Hello"}]}'

Models and pricing

qwen2.5-7b-instructqwen2.5-3b-instruct
2
Models available in preview
$0.20 / $0.60
Price per 1M tokens · input / output
Usage based
Pay for the tokens you use

Built to keep requests moving

Lower-cost routing

Live electricity prices help JouleCloud favor lower-cost power.

Automatic GPU checks

Requests go to available GPUs and can retry elsewhere if one fails.

Built for teams

API keys, budgets, and usage stay inside your organization.

DocsPricingModelsCost calculatorStatus
Private preview

Start routing AI work with live electricity prices.

Read the API docs