Skip to main content

Service mapping

This guide maps every service on the Crusoe AI Platform to its equivalent on AWS, Google Cloud (GCP), and Microsoft Azure. Use this mapping to translate familiar concepts and accelerate your deployment on our platform. For cloud-specific migration details, see the dedicated guides: AWS, GCP, and Azure.

Unified Agent & Serverless Architecture

Crusoe AI Platform provides a unified, developer-first architecture where agent, serverless, messaging, and data services work together natively. Workloads deploy with automatic HTTPS endpoints, zero complex networking, built-in session memory, and transparent project-based access controls.

Service Comparison Matrix

Crusoe AI PlatformAWSGCPAzureKey Advantages
Agents — Serverless AI agent deploymentsAmazon Bedrock AgentCore / Bedrock AgentsVertex AI Agent Engine / Gemini Enterprise Agent PlatformAzure AI Foundry Agent ServiceRuns unmodified ADK, LangGraph, or CrewAI code in a single serverless deployment model with native session state and long-term memory.
Serverless — Scale-to-zero container servicesAWS App Runner / AWS Fargate / Amazon ECSCloud RunAzure Container AppsContainer-in, scale-to-zero HTTPS endpoint out with automated SSL certificates, custom domain routing, and revision management.
Functions — Event-driven serverless functionsAWS LambdaCloud Run functionsAzure FunctionsSimple handle(event) handler model with automatic HTTPS endpoints—no API Gateway configuration required. Supports Python, Node.js, Go, and Ruby.
VectorDB — High-performance vector databaseAmazon OpenSearch Service (Vector Engine)Vertex AI Vector SearchAzure AI SearchDedicated vector database utilizing HNSW indexing with native payload filtering, avoiding the operational overhead of search-engine wrappers.
MemoryStore — Managed in-memory key-value storeAmazon ElastiCache / Amazon MemoryDBMemorystore for Redis / ValkeyAzure Managed RedisFully managed key-value store using standard Redis wire protocol; compatible with existing standard Redis client libraries without code modifications.
Pub/Sub — Distributed messaging and eventsAmazon SNS + SQS + EventBridgeCloud Pub/SubAzure Service Bus + Event GridUnified messaging service providing both topic fan-out and queue subscription patterns without requiring separate messaging products.
Code Sandbox — Isolated container executionBedrock AgentCore Code InterpreterGemini Code ExecutionAzure Container Apps Dynamic SessionsEphemeral, secure isolated execution environment created per invocation and destroyed immediately after completion.
Secrets Manager — Secure credentials managementAWS Secrets ManagerGCP Secret ManagerAzure Key VaultWrite-only project-scoped secret storage injected directly as workload environment variables, with full audit logging.
Sessions — Short-term conversation historyBedrock AgentCore Memory (short-term)Vertex AI SessionsAzure AI Foundry ThreadsNatively integrated into every agent endpoint—conversation history is automatically managed and replayed without external state infrastructure.
Memory — Long-term agent memory storageBedrock Knowledge Bases / AgentCore Long-term MemoryVertex AI Memory BankAzure AI Search GroundingBuilt-in long-term memory store allowing agents to memorize conversation contexts and perform semantic search across invocations via native tools.
MCP Servers — Model Context Protocol endpointsBedrock AgentCore GatewayCustom MCP HostingCustom MCP HostingTurnkey hosting for Python Model Context Protocol (MCP) tool endpoints with immutable versioning and single-command rollbacks.
Projects and Access — Identity and permissionsAWS IAM + AWS OrganizationsGCP Projects + IAMAzure Subscriptions + Entra ID RBACStreamlined project-based isolation with simplified admin and member roles, eliminating complex IAM policy documents.
Service Accounts & Keys — Machine identitiesIAM Roles + Access KeysGCP Service Accounts + KeysEntra Service Principals & Managed IdentitiesScoped machine identities belonging directly to projects, producing single-reveal API keys for secure automation and CI/CD pipelines.
Quotas and Audit — Resource tracking & security logsAWS Service Quotas + CloudTrailGCP Quotas + Cloud Audit LogsAzure Quotas + Activity LogReal-time resource quota tracking and always-on project audit logging included out of the box.
Embed Chat — Turnkey web chat widgetCustom Frontend WidgetCustom Frontend WidgetCopilot Studio Embed / Custom WidgetTurnkey web chat embed widget configurable with a single public key and domain allowlist.

Architectural Advantages

When evaluating the Crusoe AI Platform alongside traditional cloud providers, key architectural advantages stand out:

  • Single Generation & Purpose-Built Services: Avoid navigating retiring product generations or fragmented service choices. Every service on the platform is purpose-built and modern.
  • Zero API Gateway Configuration: Functions and services automatically receive a secure HTTPS address on deployment. You do not need to configure separate API gateways, routes, or custom domain bindings.
  • Unified Agent & Developer Experience: AI agents, serverless functions, containerized services, and data stores share a unified CLI (platformctl), single authentication model, and integrated project environment.
  • Structural Data Privacy: Provider administration operates under strict break-glass protocols. Platform staff hold no standing access to tenant data, and all access requests are logged to your immutable audit log.
  • Predictable Resource Quotas: Capacity is managed via transparent project quotas visible in real time, avoiding complex billing sub-SKUs and multi-tiered resource metering.

Next Steps