Your agent can write code. Now it can run it too.
Try FreeAgentCloud gives AI agents on-demand compute through a simple API. Write code, send it, get results back. No VPS setup, no SSH keys, no DevOps.
Every execution runs in a sandboxed container with no network access and strict resource limits. Safe by default.
Pay per minute of compute. Minimum 1 minute. No subscriptions.
Example: Run Python for 30 minutes on micro tier = $0.003
Test endpoints are free with rate limits. No payment required.
curl -X POST https://agent-runner.com/test/run \
-H "Content-Type: application/json" \
-d '{"runtime": "python", "code": "print(sum(range(100)))"}'
curl -X POST https://agent-runner.com/test/compile \
-H "Content-Type: application/json" \
-d '{"language": "rust", "code": "fn main() { println!(\"Hello!\"); }"}'
{
"success": true,
"stdout": "4950",
"stderr": "",
"exit_code": 0,
"duration_ms": 125,
"receipt": {
"input_hash": "6d652de8...",
"output_hash": "85d84cd8...",
"combined_hash": "5c47b143...",
"signature": "OvQHMa5l...",
"signer": "GYn41mEKDRmvhKuDJz1VV9tbdYzvXnHeedANPg9DLehL"
}
}
Every response includes a cryptographically signed receipt proving what code ran and what output was produced.
Use this to prove "I ran this code and got this result" - useful for auditing, debugging, or dispute resolution.
For paid endpoints, use the x402 protocol with USDC on Solana.
POST /run
{"runtime": "python", "code": "...", "tier": "small", "max_minutes": 10}
{
"x402Version": 1,
"accepts": [{
"scheme": "exact",
"network": "solana-mainnet",
"maxAmountRequired": "2000", // 0.002 USDC (in micro-units)
"payTo": "GYn41mEKDRmvhKuDJz1VV9tbdYzvXnHeedANPg9DLehL",
"asset": "EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v"
}]
}
POST /run
X-Payment: <base64-encoded payment data>
{"runtime": "python", "code": "...", "tier": "small", "max_minutes": 10}
Fallback: http://131.153.243.233:3000
Health check. Returns {"status": "ok"}
Get current pricing tiers and supported runtimes.
Free test execution (10 second limit, micro tier).
{"runtime": "python|node|bash|rust", "code": "...", "stdin": "optional"}
Free test compilation (2 minute limit, small tier).
{"language": "rust|rust-wasm|c|zig", "code": "..."}
Paid execution. Requires x402 payment.
{"runtime": "...", "code": "...", "tier": "micro|small|medium|large|xlarge", "max_minutes": 1-60}
Paid compilation. Returns base64-encoded binary. Requires x402 payment.
{"language": "...", "code": "...", "tier": "...", "max_minutes": 1-60}
If you're building an AI agent that needs to execute code, AgentCloud handles the infrastructure so you don't have to.
Fund your agent's Solana wallet with USDC and let it run code whenever it needs to.
You can write code. Now you can run it too.
Because agents shouldn't have to ask permission to run code.