The unified interface for compute

Instantly give your model access to a unified CLI supporting OpenRouter, SF Compute, Modal, Vast.ai, Lambda, Hyperbolic, and more - without separate API keys, authentication, and quotas. Just deposit credits, and we’ll handle everything else with no platform fees. Try for free with $1 of credit with no card required.

Programmatically giving each agent its own wallet to spin up compute is now as simple as…

Creating a virtual wallet:

spin wallet create --name experiment --budget 25
spin_agent_wal_72f8…

With a single command, your agent can spin up compute using that wallet:

export SPIN_BROKER_TOKEN=spin_agent_wal_72f8…
spin compute create --provider vastai --spec @lease.json --request-id gpu-001
8f3a
# test the gpus
spin compute exec 8f3a 'nvidia-smi'

It can also call any model on OpenRouter with the same token, billed to the wallet:

curl https://api.spin.inductionlabs.com/v1/chat/completions \
    -H "Authorization: Bearer $SPIN_BROKER_TOKEN" \
    -H "Content-Type: application/json" \
    -H "Idempotency-Key: reply-001" \
    -d '{
      "model": "openai/gpt-6-astra",
      "messages": [{"role": "user", "content": "Hello"}]
    }'

For more info, read the documentation or the API reference.