Send a prompt or code. Get results instantly. Python, JavaScript, TypeScript — sandboxed, secure, with your secrets injected safely.
# Generate + Execute with one API call curl -X POST https://py-api.com/api/v1/generate \ -H "x-api-key: pyapi_..." \ -d '{ "prompt": "Scrape top 5 Hacker News stories", "language": "python" }' { "code": "import requests...", "output": "1. Show HN: ...\n2. ...", "exitCode": 0 }
Describe what you need in plain English. Claude generates production-ready code and executes it instantly.
Server never sees your code. Argon2id + XChaCha20-Poly1305. Decryption only inside the sandbox. Same approach as Signal & Apple.
Python, JavaScript, TypeScript. Each execution runs in an isolated Firecracker microVM sandbox.
Install any pip or npm package. Set persistent packages or add per-request. Up to 50 packages per user.
Every execution and secret access logged with timestamps and IPs. SOC 2 and ISO 27001 ready.
Per-user rate limiting, daily execution quotas, and per-request timeouts. Built for production workloads.
One POST request. You get a JWT token and a persistent API key (pyapi_...) for all future requests.
API keys, database passwords, tokens — encrypted with AES-256-GCM and injected as environment variables when your code runs.
Send code to /execute or a natural language prompt to /generate. Results in seconds, fully sandboxed.
cURL, Python requests, Node.js fetch, Postman — or any language with HTTP support.
curl -X POST https://py-api.com/api/v1/execute \ -H "x-api-key: pyapi_abc123" \ -H "Content-Type: application/json" \ -d '{ "code": "import numpy as np\nprint(np.random.randn(5))", "language": "python", "packages": ["numpy"] }'
curl -X POST https://py-api.com/api/v1/execute \ -H "x-api-key: pyapi_abc123" \ -d '{ "code": "const r = await fetch(process.env.API_URL); console.log(await r.json())", "language": "javascript" }'
curl -X PUT https://py-api.com/api/v1/secrets/OPENAI_API_KEY \ -H "x-api-key: pyapi_abc123" \ -d '{ "value": "sk-proj-...", "description": "OpenAI production key" }' // Now use it in any execution: import os key = os.environ["OPENAI_API_KEY"]
Encrypt client-side. We store an opaque blob. Decryption happens only inside the isolated sandbox. Our server never sees your source code.
Argon2id KDF (128MiB, GPU-resistant). Your password derives the encryption key. Never stored on our servers.
NIST FIPS 203 post-quantum standard. Resistant to quantum computers. Same as Chrome TLS and AWS KMS.
24-byte nonces eliminate collision risk. Constant-time by design. Used by Signal, Cloudflare, and WireGuard.
# 1. Encrypt code locally (or use /zk/encrypt helper) curl -X POST https://py-api.com/api/v1/zk/encrypt \ -H "x-api-key: pyapi_..." \ -d '{"masterPassword": "MyP@ss!", "code": "print(42)"}' {"encryptedPayload": {"v":1, "salt":"...", "nonce":"...", "ciphertext":"..."}} # 2. Store encrypted blob (server can't read it) curl -X PUT https://py-api.com/api/v1/zk/code/my-script \ -H "x-api-key: pyapi_..." \ -d '{"encryptedPayload": {...}}' # 3. Execute — decryption inside sandbox only curl -X POST https://py-api.com/api/v1/zk/execute/my-script \ -H "x-api-key: pyapi_..." \ -d '{"masterPassword": "MyP@ss!"}' {"output": "42", "zeroKnowledge": true}
Start free. Scale as you grow.
Register, get your API key, and run your first execution. No credit card required.