Environments
Isolated, customizable execution contexts for your AI agents. Configure runtimes, packages, and tools in secure containers with persistent cloud storage.
Your Agents' Workspace
Every environment comes with its own persistent file storage — a dedicated workspace where your agents can read, write, and organize files. Think of it as giving your AI assistant its own desk with drawers that stay organized between conversations.
Work Continues Across Sessions
Your files persist between conversations. Pick up exactly where you left off — no need to re-upload or re-explain.
Agents Save Their Work
Code, reports, images — everything agents create is saved automatically. Access results anytime from any device.
Ready When You Are
Pre-load documents, datasets, or code. Your agents have instant access to everything they need to get started.
Supported Runtimes
Plus Java, Ruby, and PHP. Each runtime version can be configured per environment.
Features
Isolated Containers
Each environment runs in its own secure Docker container with dedicated resources and complete isolation from other workloads.
Multi-Language Support
Run Python, Node.js, Go, Java, Ruby, Rust, and PHP with configurable versions. Switch between Python 3.9-3.13 or Node 18-22 with a click.
Package Management
Install any system packages (apt), Python packages (pip), or Node packages (npm) without conflicts. Each environment maintains its own dependencies.
Persistent Cloud Storage
Workspaces are backed by Google Cloud Storage for durability. Your files persist across sessions and can sync with local development.
Terminal Access
Get direct terminal access to your running environments. Debug issues, run commands, and interact with your workspace in real-time.
MCP Server Integration
Extend agent capabilities with Model Context Protocol servers. Add custom tools, databases, and external services to any environment.
Encrypted Secrets
Store API keys and sensitive data securely. Environment variables and secrets are encrypted at rest and only accessible within your environment.
Fast Startup
Warm container pools ensure sub-second startup times. Your agents start executing immediately without waiting for container provisioning.
How It Works
Configure
Choose your runtime versions, install packages, set environment variables, and configure MCP servers.
Build
We automatically build a Docker image with your configuration. Images are cached for instant startup.
Execute
Assign the environment to your agents. They execute in isolated containers with full access to your configuration.
Use Cases
Development & Testing
Create isolated environments for each project with specific dependencies. Test code changes without affecting production.
Data Processing
Set up environments with pandas, numpy, and other data science packages pre-installed for efficient data analysis.
Web Scraping
Configure environments with browser automation tools, proxies, and parsing libraries for reliable web data extraction.
API Integrations
Build environments with MCP servers connected to your databases, CRMs, and internal tools for seamless automation.
Enterprise-Grade Security
Container Isolation
Each environment runs in its own isolated container with no access to other workloads.
Encrypted Storage
All data at rest is encrypted. Secrets are stored using AES-256-GCM encryption.
Network Control
Configure internet access per environment. Restrict outbound connections when needed.
Ready to get started?
Create your first environment and start running agents in isolated, customizable containers.