Code execution
Routed across E2B and Daytona (identically priced — routing here is about capacity failover and latency, not cost). Each call gets a fresh isolated sandbox, torn down afterward.
Request
Section titled “Request”POST /v1/tools/code{ "language": "python", "code": "import math\nprint(math.pi ** 2)", "timeout_seconds": 60, "routing": { "sort": "latency" }}| Field | Type | Notes |
|---|---|---|
language | enum | python (default) | javascript |
code | string | ≤ 200 KB |
timeout_seconds | int | 1–300, default 60 |
Result
Section titled “Result”{ "result": { "stdout": "9.869604401089358\n", "stderr": "", "exit_code": 0, "duration_ms": 1840 }}Billing is per vCPU-second (usage.units).