Secure, isolated sandboxes where AI agents run code, browse the web, and ship to production — all from one API.
Machines, tools, and networking — raw primitives with zero abstraction tax.
Run Python, Node.js, Bash, and more in isolated containers with full stdin/stdout streaming.
Headless Chrome instances for web scraping, testing, and autonomous browsing tasks.
Filesystem persistence across sessions. Agents can read, write, and manage files like a real workstation.
WebSocket-based output streaming. Watch your agents think and execute in real-time.
Fine-grained network policies per sandbox. Allow, deny, or proxy specific endpoints.
Run agents for multiple customers with full resource and data isolation between tenants.
Full Linux sandboxes with filesystem, networking, and process isolation. Ephemeral for one-off tasks, persistent for agents that need a home.
Shell execution, file I/O, browser automation, web search — raw primitives your agents compose into autonomous workflows.
Turn any sandbox into a live deployment with one API call. Custom domains, SSL, and edge routing — from prototype to production instantly.
Your agent deserves more than an API call. Give it a living environment — a sandbox that boots, runs, persists, connects, and ships to the world.
Isolated sandbox per agent — always on, always ready.
Code, browse, files, network — one SDK, zero setup.
Deploy with custom domains, SSL, and routing in one call.
Logs, metrics, and SSH access — full visibility into every agent session.
Real use cases. Real infrastructure. Agents that don't just reply — they operate.
Run any agent — securely in the cloud
Deploy open-source agents like OpenClaw, AutoGPT, or your own — each gets a secure, isolated cloud sandbox with full power. Persistent memory, unrestricted networking, browser access, and real compute. Your agent runs 24/7 without touching your local machine.
Full SDKs for Python, JavaScript, and Go — or deploy agents straight from the dashboard. Your call.
from oblien import Sandbox
# Create a persistent sandbox for your agent
sandbox = Sandbox.create(persistent=True, memory="4GB")
# Your agent runs code, browses, stores files
result = sandbox.exec("python agent.py --task research")
sandbox.fs.write("/data/report.md", result.output)
# Deploy to a live URL when ready
url = sandbox.publish(port=3000, domain="my-agent.oblien.app")
# Or just SSH in to inspect
sandbox.ssh() # → root@sandbox-3d1b:~$Deploy any open-source agent in seconds — no code required.
Not a restricted API. Not a thin wrapper. A full Linux system with shell, filesystem, network, and browser — all programmable through one unified SDK.
Real Linux with root. Install anything, run services, manage processes.
Read, write, manage files. Persistent volumes that survive restarts.
Ship instantly. One call to publish any port to a live URL with auto SSL.
Granular network control per sandbox. Allow/block domains, IPs, ports.
Headless browser with rotating proxies. Scrape and extract at scale.
Real Linux with root. Install anything, run services, manage processes.
Read, write, manage files. Persistent volumes that survive restarts.
Ship instantly. One call to publish any port to a live URL with auto SSL.
Granular network control per sandbox. Allow/block domains, IPs, ports.
Headless browser with rotating proxies. Scrape and extract at scale.
Web search and content extraction. Search, extract, and process in one call.
WebSocket connections for live output. No polling, no delays.
Outbound requests, run servers, expose ports. Full networking stack.
Multiple sandboxes in parallel. Spawn sub-agents, share data, coordinate.
Real Linux with root. Install anything, run services, manage processes.
Read, write, manage files. Persistent volumes that survive restarts.
Ship instantly. One call to publish any port to a live URL with auto SSL.
Granular network control per sandbox. Allow/block domains, IPs, ports.
Web search and content extraction. Search, extract, and process in one call.
WebSocket connections for live output. No polling, no delays.
Outbound requests, run servers, expose ports. Full networking stack.
Multiple sandboxes in parallel. Spawn sub-agents, share data, coordinate.
Real Linux with root. Install anything, run services, manage processes.
Read, write, manage files. Persistent volumes that survive restarts.
Ship instantly. One call to publish any port to a live URL with auto SSL.
Granular network control per sandbox. Allow/block domains, IPs, ports.
Everything a developer would do on their own machine — now available to your agent via a single API call.
Temporary for a task. Persistent for an agent. Deployed to the world. Every sandbox adapts to what your agent needs right now.
Temporary sandboxes for one-off tasks. Your agent gets a fresh environment, executes its job, and the sandbox self-destructs. Zero cleanup, zero cost when idle.
Most teams waste months building sandbox infra. We ship it as one import.
.create().exec().tools.*.deploy().ssh().namespaceStart free, scale as you grow. No hidden fees, no surprise bills. Every plan includes full access to all APIs and tools.
For experimenting and building prototypes
For shipping production agents
For teams running agents at scale
For organizations with custom needs
Every namespace automatically tracks per-user usage, enforces limits, and calculates costs. No billing system to build. No user management layer to maintain.
Every API call is automatically attributed to the right user. No instrumentation needed — just pass a user ID.
Every resource maps to credits — compute, browser, search, storage. Set limits per user, charge what you want.
Set sandbox hours, concurrent sandboxes, API rate limits — all per user. Enforce automatically at the platform level.
Bring your own auth. Pass any user identifier — our namespace system handles isolation, tracking, and billing from there.
const ns = await oblien.namespace.create({
userId: "user_abc123",
limits: {
sandboxHours: 10,
credits: 500
}
});
// All usage auto-tracked
const sandbox = await ns.sandbox.create();
// → credits deducted automatically
// → limits enforced in real-timeEvery account starts with free credits — enough to prototype, test, and ship your first agent. No credit card required. When you're ready to scale, credits refill automatically based on your plan.
Sandboxes, execution, tools, and deployment. Everything your agents need to go from prototype to production.
Get updates on new primitives, sandbox features, and best practices for building production-grade AI agents.
No spam. Unsubscribe at any time.