Core Concepts
Understanding these core concepts will help you make the most of AI Computer's capabilities and build secure, efficient applications.
Sandboxes
Sandboxes are isolated environments where your Python code runs securely. Each sandbox:
- Is completely isolated from other sandboxes and the host system
- Has its own filesystem and process space
- Automatically cleans up after code execution
- Has configurable resource limits
Execution Models
AI Computer supports two main execution models:
Synchronous Execution
Simple request-response model where code is executed and the complete output is returned when finished.
Streaming Execution
Real-time output streaming for long-running processes or interactive applications.
Security
Security is at the core of AI Computer's design. Our security model includes:
- Process-level isolation using containerization
- Resource limits and quotas
- Network access controls
- Filesystem isolation
- Automatic cleanup of sandbox environments