AI Integration

Encore Cloud supercharges AI-powered development by letting AI agents provision real infrastructure in your AWS or GCP account with automatic guardrails.

When you connect your cloud account to Encore Cloud, AI-generated code that declares databases, pub/sub topics, cron jobs, and other primitives gets automatically provisioned with production-ready defaults: proper networking, IAM permissions, and security configurations.

How It Works

AI writes infrastructure code using Encore's declarative primitives
Push to GitHub to trigger a deployment
Encore Cloud provisions the infrastructure in your AWS/GCP account with automatic guardrails
Preview environments let you test AI-generated changes in isolation

This enables fast iterative development: AI generates code, you push it and validate in preview environments, then deploy seamlessly with automatic infrastructure provisioning.

Infrastructure with Guardrails

When AI declares infrastructure using Encore primitives, Encore Cloud provisions it in your cloud with automatic guardrails:

Databases: Proper networking, encryption at rest, automated backups
Pub/Sub: Dead letter queues, retry policies, proper IAM roles
Secrets: Encrypted storage, access controls
Services: Load balancing, health checks, auto-scaling

AI doesn't need to know the intricacies of AWS or GCP. It just declares what it needs, and Encore Cloud handles the cloud-specific configuration.

You stay in control: review infrastructure changes in pull requests, approve or deny resource additions, and use infrastructure configuration to customize defaults per environment.

Preview Environments

Preview environments are perfect for testing AI-generated changes. Each pull request gets its own isolated environment with real infrastructure.

This means you can:

Let AI generate features and immediately test them with real databases and services
Review AI-generated infrastructure changes before they hit production
Catch issues early in isolated environments

Production Observability

Encore Cloud provides distributed tracing and metrics across all your environments. You can:

Analyze traces to debug issues across services
Inspect timing to find bottlenecks
Compare behavior between preview and production environments

MCP Servers

Encore provides Model Context Protocol (MCP) servers that give AI agents rich context about your application. MCP is an open standard that lets large language models access contextual information from external sources.

Local MCP Server

The Local MCP Server runs alongside encore run and gives AI tools access to your local application's metadata, traces, databases, source code, and more. It's ideal for connecting AI assistants like Cursor or Claude Desktop to your local development environment.

Cloud MCP Server

The Cloud MCP Server gives AI agents access to your deployed environments on Encore Cloud, including production traces, deployment status, environment configuration, and infrastructure metadata. It's hosted at https://api.encore.cloud/mcp and authenticates via OAuth or a Bearer API token.

Connecting Your Cloud

To deploy your Encore app to your cloud:

Encore Cloud provisions infrastructure in your cloud account based on the primitives declared in your code. You maintain full control and ownership of your infrastructure.

What AI Can Provision

With Encore Cloud connected, AI-generated code can provision:

Resource	AWS	GCP
Databases	RDS (PostgreSQL)	Cloud SQL
Pub/Sub	SNS + SQS	Cloud Pub/Sub
Object Storage	S3	Cloud Storage
Cron Jobs	CloudWatch Events	Cloud Scheduler
Secrets	Secrets Manager	Secret Manager
Caching	ElastiCache	Memorystore

All resources are provisioned with security best practices like least-privilege IAM policies, private networking, and encryption. See the AWS and GCP infrastructure docs for specifics.