What is the Crusoe AI Platform?
The Crusoe AI Platform is a fully managed cloud for building, deploying, and scaling AI agents, serverless functions, and containerized microservices.
You write your code—whether it is a simple function or a complex multi-agent system—and deploy it with a single command or click. The platform automatically handles hosting, security, scaling, state management, and infrastructure so you can focus on building products.
Who It's For
- Agent Developers: Deploy unmodified open-source framework code (ADK, LangGraph, CrewAI) with built-in memory, tools, and session management.
- Software Engineers: Run serverless functions and containerized microservices without provisioning, scaling, or operating any infrastructure yourself.
- Engineering Teams: Collaborate in isolated project environments with role-based access control (RBAC), API key management, usage quotas, and complete audit logging.
How It Works
Everything you build lives inside a Project, which is an isolated workspace for your team.
You interact with the platform through two primary interfaces:
- Web Console: An intuitive web interface located at console.codyhill.dev for visual management, monitoring, testing, and team collaboration.
- CLI (
platformctl): A command-line tool designed for local development, scripting, and continuous integration/continuous deployment (CI/CD) pipelines.
Automatic Scaling
The Crusoe AI Platform automatically scales your workloads based on incoming traffic:
- Scale to Zero: When your service is not receiving requests, it scales down to zero running instances so you incur no compute costs.
- Instant Warm-Up: When a request arrives, the platform provisions an instance in milliseconds to handle the payload.
- Autoscaling: As demand increases, the platform automatically provisions additional instances to maintain high performance.
Service Directory
| Service | Category | Description | Documentation |
|---|---|---|---|
| Agents | Compute | Deploy AI agents with session state, long-term memory, and interactive chat endpoints. | Agent Guide |
| Functions | Compute | Event-driven serverless code execution for Python, Node.js, Go, and Ruby. | Function Guide |
| Serverless Containers | Compute | Deploy custom container images that scale automatically based on HTTP traffic. | Serverless Guide |
| API Gateway | Networking | Secure your endpoints with custom domains, rate limits, API keys, and IP allowlists. | Gateway Guide |
| VectorDB | Data | Managed vector database for semantic search, document retrieval, and RAG pipelines. | VectorDB Guide |
| MemoryStore | Data | Redis-compatible, high-performance in-memory cache for session persistence. | MemoryStore Guide |
| Pub/Sub | Messaging | Real-time event streaming and message queuing for asynchronous microservices. | Pub/Sub Guide |
| Secrets Manager | Security | Centralized, encrypted secret storage with automatic environment variable injection. | Secrets Guide |
| Code Sandbox | Security | Secure, isolated sandbox environments for executing untrusted user code safely. | Sandbox Guide |
| MCP Servers | Integration | Host Model Context Protocol (MCP) endpoints to share tools across agents and clients. | MCP Guide |
Next Steps
- Create an Account: Sign in and set up your first organization and project.
- Install the CLI: Set up
platformctlon your workstation. - Deploy Your First Agent: Build and deploy a live AI agent in under 5 minutes.
- Deploy Your First Function: Ship a serverless function endpoint.
- Explore Core Concepts: Learn about projects, access control, and platform quotas.