vaaya / tools / modal / sandbox-create

Modal `sandbox-create` for AI agents

Name: Modal sandbox-create for AI agents
Brand: Modal

Modal — Create a sandboxed compute environment (CPU by default, 300s timeout). Pass `gpu` (T4 | L4 | A10G | A100 | H100) for GPU workloads like model inference.

varies by params per call

npx @vaaya/mcp install

Parameters

Param	Type	What
`command`	`string \| string[]`
`gpu`	`"T4" \| "L4" \| "A10G" \| "A100" \| "H100"`
`image`	`string`
`timeout`	`integer`

Call it

From any MCP client with Vaaya connected, one use call:

{
  "service": "modal",
  "action": "sandbox-create",
  "params": {},
  "max_cost_cents": 100
}

Or over HTTP with a Vaaya API key:

curl -X POST https://vaaya.ai/api/run/modal/sandbox-create \
  -H "Authorization: Bearer $VAAYA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"max_cost_cents":100}'

Notes

Settled via MPP (Tempo chain) — no API key needed, the payment IS the auth. Sandbox lives for up to 300s by default. Pair with `modal:sandbox-status` to check state and `modal:sandbox-terminate` to clean up.

The facts

Call

modal/sandbox-create

Price

varies by params — the listed figure is a cap; you pay what the provider settles. Charged only on success.

Pillar

Compute

Provider

Modal · modal.com

Settlement

Paid per call over the Machine Payments Protocol from Vaaya’s treasury; you are debited the settled amount.

Credentials

Held server-side by Vaaya; your agent never receives a Modal key.

Give your agents this call.

Connect Vaaya once, then describe the outcome or call it directly. Paid per call, from one balance.