cli

name	cli
description	Trebuchet CLI commands for initializing projects, deploying actors, checking status, local development, and infrastructure management. Use when users need to deploy, configure, or manage Trebuchet projects from the command line.

Trebuchet CLI

Command-line interface for deploying and managing Trebuchet distributed actors.

Overview

The trebuchet CLI provides tools for:

• Initializing new projects
• Deploying actors to cloud platforms
• Managing deployments
• Running local development servers
• Generating infrastructure configuration
• Automatic Xcode project support (since v0.4.0)
• Intelligent dependency analysis for minimal deployments

Xcode Project Support

Zero configuration required! The CLI automatically detects and works with:

• Xcode projects (.xcodeproj)
• Xcode workspaces (.xcworkspace)
• Swift Package Manager projects (Package.swift)

Automatic Detection

The CLI automatically:

1. Detects project type (Xcode vs SPM)
2. Extracts package names from Package.swift when available
3. Generates appropriate package manifests
4. Adapts behavior across all commands (dev, deploy, generate server)

Automatic Dependency Analysis

The CLI uses SwiftSyntax to intelligently analyze actor dependencies:

What it analyzes:

• Actor method signatures (parameters and return types)
• Transitive dependencies (e.g., PlayerInfo → GameStatus)
• Complex types: generics (Array<T>), optionals (T?), nested types
• Only types actually used by actors (symbol-scoped analysis)

What it filters out:

• Standard library types (String, Int, UUID, etc.)
• Unrelated types in the same file as dependencies
• Files that actors don't depend on

Result: Only 1-5 files typically copied vs. entire app

Cascade Prevention

The CLI prevents copying your entire codebase:

• Symbol-level analysis: Only analyzes types actually used by actors
• Smart filtering: Doesn't cascade through unrelated dependencies
• 20-60x improvement over naive file-level analysis

Example: If your actor uses PlayerInfo from Models.swift, but Models.swift also contains UnrelatedType, the CLI won't cascade to UnrelatedType's dependencies.

Works Everywhere

Xcode support works across all commands:

# Local development - zero config
trebuchet dev

# Generate server package - automatic dependency copying
trebuchet generate server --output ./my-server

# Deploy to cloud - works with Xcode projects
trebuchet deploy --provider aws

Installation

Using Mint (Recommended)

mint install briannadoubt/Trebuchet

Build from Source

git clone https://github.com/briannadoubt/Trebuchet.git
cd Trebuchet
swift build -c release --product trebuchet
cp .build/release/trebuchet /usr/local/bin/

Verify Installation

trebuchet --version

Commands

trebuchet init

Initialize a new Trebuchet project configuration.

trebuchet init --name my-project --provider aws

Options:

• --name (required): Project name
• --provider: Cloud provider (aws, fly)
• --region: Default deployment region
• --output: Output file path (default: trebuchet.yaml)

Example:

trebuchet init --name game-server --provider aws --region us-east-1

Creates trebuchet.yaml:

name: game-server
version: "1"

defaults:
  provider: aws
  region: us-east-1
  memory: 512
  timeout: 30

actors: {}
state:
  type: dynamodb
discovery:
  type: cloudmap

trebuchet deploy

Deploy actors to the cloud.

trebuchet deploy --provider aws --region us-east-1

Options:

• --provider: Cloud provider (aws, fly)
• --region: Deployment region
• --environment: Environment name (production, staging, etc.)
• --dry-run: Preview deployment without making changes
• --verbose: Show detailed output
• --config: Path to config file (default: trebuchet.yaml)

Examples:

# Deploy to AWS in us-east-1
trebuchet deploy --provider aws --region us-east-1

# Deploy to Fly.io
trebuchet deploy --provider fly --region iad

# Dry run to preview changes
trebuchet deploy --dry-run --verbose

# Deploy to staging environment
trebuchet deploy --environment staging

# Use custom config file
trebuchet deploy --config custom.yaml

Output:

Discovering actors...
  ✓ GameRoom (Sources/Actors/GameRoom.swift)
  ✓ Lobby (Sources/Actors/Lobby.swift)

Building for Lambda (arm64)...
  Building executable...
  Packaging bootstrap...
  ✓ Package built (14.2 MB)

Deploying to AWS...
  Creating Lambda function...
  ✓ Lambda: arn:aws:lambda:us-east-1:123:function:game-server
  Creating API Gateway...
  ✓ API Gateway: https://abc123.execute-api.us-east-1.amazonaws.com
  Creating DynamoDB table...
  ✓ DynamoDB: game-server-actor-state
  Creating CloudMap namespace...
  ✓ CloudMap: game-server namespace

Deployment complete!
Endpoint: https://abc123.execute-api.us-east-1.amazonaws.com

trebuchet status

Check the status of deployed actors.

trebuchet status

Options:

• --verbose: Show detailed information
• --provider: Filter by provider
• --config: Path to config file

Example:

trebuchet status --verbose

Output:

Project: game-server
Provider: aws
Region: us-east-1

Actors:
  GameRoom
    Status: Active
    Memory: 1024 MB
    Timeout: 60s
    Last Update: 2026-01-27 10:30:00 UTC
    Invocations (24h): 1,234

  Lobby
    Status: Active
    Memory: 256 MB
    Timeout: 30s
    Last Update: 2026-01-27 10:30:00 UTC
    Invocations (24h): 567

Infrastructure:
  Lambda: arn:aws:lambda:us-east-1:123:function:game-server
  API Gateway: https://abc123.execute-api.us-east-1.amazonaws.com
  DynamoDB: game-server-actor-state
  CloudMap: game-server

trebuchet undeploy

Remove deployed infrastructure.

trebuchet undeploy

Options:

• --provider: Cloud provider
• --region: Deployment region
• --force: Skip confirmation prompt
• --config: Path to config file

Example:

trebuchet undeploy --force

Output:

⚠️  This will delete all infrastructure and data!

Removing infrastructure...
  ✓ Lambda function deleted
  ✓ API Gateway deleted
  ✓ DynamoDB table deleted
  ✓ CloudMap namespace deleted

Undeployment complete.

trebuchet dev

Run actors locally for development.

trebuchet dev --port 8080

Options:

• --port / -p: HTTP server port (default: 8080)
• --host / -h: Host to bind to (default: localhost)
• --verbose / -v: Enable verbose build output

Examples:

# Start on default port (8080)
trebuchet dev

# Custom port and host
trebuchet dev --port 3000 --host 0.0.0.0

# Verbose output
trebuchet dev --verbose

What it does:

1. Discovers all @Trebuchet actors in your project
2. Builds your project with swift build (skipped for Xcode projects)
3. Analyzes actor dependencies and copies only required files
4. Generates a local development runner in .trebuchet/
5. Starts an HTTP server using CloudGateway.development()
6. Exposes actors at /invoke endpoint
7. Provides health check at /health endpoint

Note: For Xcode projects, the CLI skips swift build since Xcode handles compilation. The dev server uses the generated package in .trebuchet/ instead.

Output:

Starting local development server...

Found actors:
  • GameRoom
  • Lobby

Building project...
✓ Build succeeded

Generating development server...
✓ Runner generated

Starting server on localhost:8080...

Starting local development server...

  ✓ Exposed: GameRoom
  ✓ Exposed: Lobby

Server running on http://localhost:8080
Health check: http://localhost:8080/health
Invocation endpoint: http://localhost:8080/invoke

Press Ctrl+C to stop

trebuchet generate

Generate deployment artifacts.

trebuchet generate server

Generate a standalone server package from your actors.

trebuchet generate server --output ./my-server

Options:

• --output: Output directory for generated server (default: .trebuchet/server)
• --config: Path to trebuchet.yaml
• --verbose / -v: Enable verbose output
• --force: Force regeneration even if server exists

Examples:

# Generate server in default location
trebuchet generate server

# Generate in custom directory
trebuchet generate server --output ./deploy/server

# Force regeneration
trebuchet generate server --force --verbose

What it does:

1. Discovers all @Trebuchet actors
2. Generates a standalone Swift package
3. Creates Package.swift with dependencies
4. Generates main.swift with actor bootstrapping
5. Ready to deploy or run independently

Output:

Loading configuration...

Discovering actors...
  ✓ GameRoom
  ✓ Lobby

Generating server package...
  ✓ Package manifest created
  ✓ Bootstrap code generated
  ✓ Server configured

✓ Server package generated at .trebuchet/server

To deploy:
  trebuchet deploy

To run locally:
  cd .trebuchet/server
  swift run

Configuration File (trebuchet.yaml)

Basic Structure

name: my-project
version: "1"

defaults:
  provider: aws
  region: us-east-1
  memory: 512
  timeout: 30

actors: {}
environments: {}
state: {}
discovery: {}

Actor Configuration

actors:
  GameRoom:
    memory: 1024          # Memory in MB
    timeout: 60           # Timeout in seconds
    stateful: true        # Enable state persistence
    isolated: true        # Dedicated Lambda
    environment:          # Environment variables
      LOG_LEVEL: debug
      MAX_PLAYERS: "10"

  Lobby:
    memory: 256
    timeout: 30

Environment Configuration

environments:
  production:
    region: us-west-2
    memory: 2048
    environment:
      LOG_LEVEL: warn
      ENABLE_METRICS: "true"

  staging:
    region: us-east-1
    memory: 512
    environment:
      LOG_LEVEL: debug

State Configuration

state:
  type: dynamodb
  tableName: custom-table-name    # Optional

Supported types:

• dynamodb: AWS DynamoDB (for AWS deployments)
• postgresql or postgres: PostgreSQL database (for Fly.io or custom deployments)
• memory: In-memory (development only)

Discovery Configuration

discovery:
  type: cloudmap
  namespace: my-namespace

Supported types:

• cloudmap: AWS Cloud Map
• memory: In-memory (development only)

WebSocket Configuration

websocket:
  enabled: true
  stage: production
  routes:
    - $connect
    - $disconnect
    - $default

connections:
  type: dynamodb
  table: my-app-connections

Actor Discovery

The CLI automatically discovers actors in your codebase:

// Sources/Actors/GameRoom.swift
@Trebuchet
distributed actor GameRoom {
    distributed func join(player: Player) -> RoomState
}

The CLI scans:

• Sources/ directory
• All .swift files
• Actors marked with @Trebuchet
• Actor annotations in comments

Build Process

When deploying to AWS Lambda, the CLI:

1. Detects project type (Xcode vs SPM)
2. Discovers actors using SwiftSyntax
3. Analyzes dependencies to identify required types
4. Copies actor files and dependencies (1-5 files typically)
5. Generates bootstrap code for Lambda handler
6. Builds for ARM64 using Docker
7. Packages binary with swift-lambda-runtime
8. Generates Terraform for infrastructure
9. Applies infrastructure changes
10. Uploads Lambda package

Dependency Analysis Performance

• Analysis time: 50-100ms for typical projects
• Cascade prevention: Avoids copying 200+ unnecessary files in worst case
• File copying: Only 1-5 files typically copied vs entire app

Docker Build

The CLI uses Docker to cross-compile for Lambda ARM64:

FROM swift:latest
RUN swift build -c release --arch arm64

Terraform Integration

The CLI generates Terraform configuration in .trebuchet/terraform/:

.trebuchet/
└── terraform/
    ├── main.tf
    ├── variables.tf
    ├── outputs.tf
    └── lambda.zip

Customizing Terraform

You can customize the generated Terraform:

cd .trebuchet/terraform
terraform plan -var="lambda_memory=2048"
terraform apply

Using Existing Terraform

Skip CLI-generated Terraform and use your own:

trebuchet deploy --skip-terraform

Common Workflows

Initial Setup

# 1. Initialize project
trebuchet init --name my-project --provider aws

# 2. Edit trebuchet.yaml to configure actors
vi trebuchet.yaml

# 3. Test locally
trebuchet dev --port 8080

# 4. Deploy to staging
trebuchet deploy --environment staging

# 5. Deploy to production
trebuchet deploy --environment production

Iterative Development

# 1. Run local development server
trebuchet dev --port 8080

# 2. Make changes to actors

# 3. Restart dev server to test changes
# (Ctrl+C to stop, then run trebuchet dev again)

# 4. Deploy to staging
trebuchet deploy --environment staging

# 5. Verify in staging
trebuchet status --environment staging

# 6. Deploy to production
trebuchet deploy --environment production

Debugging Deployment Issues

# Preview changes without applying
trebuchet deploy --dry-run --verbose

# Check current deployment status
trebuchet status --verbose

# View generated Terraform
cat .trebuchet/terraform/main.tf

# Manually apply Terraform
cd .trebuchet/terraform
terraform plan
terraform apply

Environment Variables

The CLI respects these environment variables:

• AWS_ACCESS_KEY_ID: AWS access key
• AWS_SECRET_ACCESS_KEY: AWS secret key
• AWS_REGION: Default AWS region
• AWS_PROFILE: AWS credentials profile
• TREBUCHET_CONFIG: Path to config file

Exit Codes

• 0: Success
• 1: General error
• 2: Configuration error
• 3: Build error
• 4: Deployment error
• 5: Connection error

Troubleshooting

Actor Discovery Fails

Problem: CLI doesn't find actors

Solutions:

• Ensure actors are marked with @Trebuchet
• Check actors are in Sources/ directory
• Verify Swift files are valid syntax
• Use --verbose flag to see discovery details

Build Fails

Problem: Docker build fails

Solutions:

• Ensure Docker is running
• Check Docker has sufficient memory (4GB+)
• Verify internet connection for package downloads
• Use --verbose to see full build output

Deployment Fails

Problem: Terraform apply fails

Solutions:

• Check AWS credentials are valid
• Verify IAM permissions
• Check region is correct
• Review .trebuchet/terraform/ for errors
• Use trebuchet deploy --dry-run first

Connection Timeout

Problem: CLI can't connect to AWS

Solutions:

• Verify internet connection
• Check AWS credentials
• Try different region
• Check firewall settings

See Also

• Cloud deployment guide for deployment concepts
• AWS Lambda guide for AWS-specific deployment
• Configuration reference for trebuchet.yaml options