Neura Market
Cloud VMs with a Built-In REST API for AI Agents logo

Cloud VMs with a Built-In REST API for AI Agents

Free
FreeFree tier
Type
Open Source
Company
Oblien
Visit Website GitHub (1 stars)

About Cloud VMs with a Built-In REST API for AI Agents

Oblien provides cloud VMs (workspaces) with a built-in REST API specifically designed for AI agents. Each workspace is a hardware-isolated VM running a Runtime API on port 9990, exposing file operations, terminal sessions, command execution, search, file watcher, and MCP skills. The API can be accessed via a gateway (workspace.oblien.com) with JWT authentication or directly between workspaces over a private network using raw connection tokens. Workspaces are network-dark by default, with firewalls blocking all inbound traffic unless explicitly allowed. Cold starts take under 130 ms, warm snapshots restore instantly, and a 99.95% uptime SLA is offered. A single TypeScript SDK provides full lifecycle control (create, start, stop, delete) and runtime access. The service includes a generous free tier with pay-as-you-go pricing, no credit card required.

Key Features

Hardware-isolated VMs for each workspace
Built-in REST API on port 9990 for file, terminal, and command execution
Gateway access with JWT authentication and direct private IP access
Network-dark by default with built-in firewall
Millisecond cold starts (under 130 ms) and instant warm snapshots
TypeScript SDK for full workspace lifecycle and runtime control
99.95% uptime SLA
Generous free tier with pay-as-you-go pricing

Pros & Cons

Pros
  • True hardware isolation (not containers) for maximum security
  • Extremely fast cold start (under 130 ms) and instant state restoration
  • Comprehensive, well-documented REST API for all runtime operations
  • Generous free tier allows experimentation without payment
  • Single SDK to manage both lifecycle and runtime interactions
  • Network isolation by default, with explicit private links for controlled access
Cons
  • Requires creating private links for direct workspace-to-workspace communication, adding setup complexity
  • No GPU support mentioned, limiting use for GPU-accelerated AI workloads
  • Pricing may become expensive at scale (no explicit rates provided)
  • Dual access modes (gateway vs direct) may have a learning curve for new users

Best For

Autonomous AI agents needing a secure, isolated runtime environmentRunning code, file operations, and terminal commands programmaticallyWorkspace-to-workspace agent communication over private networkCI/CD pipelines and automated testing in isolated VMsSecure sandboxing for executing untrusted code

FAQ

How do I access the Runtime API?
You can access it via the gateway at workspace.oblien.com with a JWT (requires enabling public_access) or directly via the workspace's private IP (10.x.x.x:9990) with a raw connection token obtained from the Oblien API.
What endpoints are available in the Runtime API?
Endpoints include file operations (list, read, write, delete, mkdir, stat, stream, transfer), terminal sessions, command execution, search, file watcher, and MCP skills. A full endpoint map is provided in the documentation.
Is each workspace isolated from others?
Yes, each workspace runs in its own hardware-isolated VM with a firewall that blocks all inbound traffic by default. Even workspaces in the same account cannot reach each other until a private link is explicitly created.
Do I need a credit card to start?
No, the service offers a generous free tier with no credit card required. Your first workspace boots in under 130 ms.