SignalWire AI Agents - Cloud Functions

Deploy SignalWire AI Agents to serverless platforms: AWS Lambda, Google Cloud Functions, and Azure Functions.

Overview

Each directory contains a complete, deployable Hello World agent with:

• SWML output for SignalWire phone integration
• SWAIG functions (say_hello, get_platform_info, echo)
• Native CLI deployment scripts

Platform Comparison

Feature	AWS Lambda	Google Cloud	Azure Functions
CLI	aws	gcloud	az
Runtime	Python 3.11	Python 3.11	Python 3.11
Entry Point	lambda_handler	main	main
Max Timeout	15 min	60 min (Gen 2)	10 min (Consumption)
Cold Starts	~1-3s	~1-2s	~2-5s
Min Memory	128 MB	128 MB	Flex
Free Tier	1M requests/mo	2M invocations/mo	1M executions/mo

CLI Setup

AWS CLI

# Install AWS CLI
# macOS
brew install awscli

# Linux
curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
unzip awscliv2.zip && sudo ./aws/install

# Windows
msiexec.exe /i https://awscli.amazonaws.com/AWSCLIV2.msi

# Configure credentials
aws configure
# Enter: AWS Access Key ID, Secret Access Key, Default region, Output format

Google Cloud CLI

# Install gcloud CLI
# macOS
brew install --cask google-cloud-sdk

# Linux/Windows - Download from:
# https://cloud.google.com/sdk/docs/install

# Authenticate
gcloud auth login

# Set project
gcloud config set project YOUR_PROJECT_ID

Azure CLI

# Install Azure CLI
# macOS
brew install azure-cli

# Linux
curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bash

# Windows
winget install -e --id Microsoft.AzureCLI

# Authenticate
az login

Quick Start

AWS Lambda

cd aws
./deploy.sh

Google Cloud Functions

cd gcloud
./deploy.sh

Azure Functions

cd azure
./deploy.sh

Directory Structure

cloud-functions/
├── README.md           # This file
├── aws/
│   ├── handler.py      # Lambda handler
│   ├── requirements.txt
│   ├── deploy.sh       # AWS CLI deployment
│   └── README.md
├── gcloud/
│   ├── main.py         # Cloud Functions entry point
│   ├── requirements.txt
│   ├── deploy.sh       # gcloud CLI deployment
│   └── README.md
└── azure/
    ├── function_app/
    │   ├── __init__.py # Azure Functions handler
    │   └── function.json
    ├── host.json
    ├── requirements.txt
    ├── deploy.sh       # Azure CLI deployment
    └── README.md

Agent Features

All three agents implement identical functionality:

SWAIG Functions

Function	Description	Parameters
say_hello	Personalized greeting	name (string, required)
get_platform_info	Platform runtime information	none
echo	Echo back a message	message (string, required)

Voice Configuration

• Language: English (en-US)
• Voice: rime.spore

Environment Variables

All platforms support these optional environment variables:

Variable	Description	Default
SWML_BASIC_AUTH_USER	Basic auth username	Auto-generated
SWML_BASIC_AUTH_PASSWORD	Basic auth password	Auto-generated

Note: If not set, the SDK automatically generates secure credentials. The deploy scripts display the credentials after deployment.

Authentication

The SignalWire Agents SDK automatically enables HTTP Basic Authentication for security. You can:

1. Let the SDK generate credentials - Secure random credentials are created automatically
2. Set your own credentials - Via environment variables before deployment:

   SWML_BASIC_AUTH_USER=myuser SWML_BASIC_AUTH_PASSWORD=mypass ./deploy.sh

Testing

Test SWML Output

# Replace with your endpoint and credentials
curl -u username:password https://your-endpoint/

Test SWAIG Function

curl -u username:password -X POST https://your-endpoint/swaig \
    -H 'Content-Type: application/json' \
    -d '{"function": "say_hello", "argument": {"parsed": [{"name": "Alice"}]}}'

Local Testing with swaig-test

# AWS Lambda
cd aws
swaig-test handler.py --simulate-serverless lambda --dump-swml

# Google Cloud Functions
cd gcloud
swaig-test main.py --simulate-serverless cloud_function --dump-swml

# Azure Functions
cd azure/function_app
swaig-test __init__.py --simulate-serverless azure_function --dump-swml

SignalWire Configuration

After deployment, configure your SignalWire phone number:

1. Log into your SignalWire Space
2. Navigate to Phone Numbers
3. Select your phone number
4. Set Handle Calls Using to SWML Script
5. Enter the endpoint URL with credentials: https://user:pass@your-endpoint/

Request Flow

┌──────────────┐      ┌─────────────────┐      ┌──────────────────┐
│  SignalWire  │──────│  API Gateway/   │──────│  Cloud Function  │
│   Platform   │      │  HTTP Trigger   │      │  (Your Agent)    │
└──────────────┘      └─────────────────┘      └──────────────────┘
       │                      │                        │
       │   POST /             │                        │
       │─────────────────────>│───────────────────────>│
       │                      │     Returns SWML       │
       │<─────────────────────│<───────────────────────│
       │                      │                        │
       │   POST /swaig        │                        │
       │─────────────────────>│───────────────────────>│
       │                      │  Execute SWAIG func    │
       │<─────────────────────│<───────────────────────│
       │                      │                        │

Best Practices

Cold Start Optimization

• Initialize agent outside the handler function
• Keep dependencies minimal
• Use provisioned concurrency (AWS) or min instances (GCP/Azure)

Security

• Use environment variables for secrets
• Enable basic auth for production (enabled by default)
• Use HTTPS endpoints only

Monitoring

• Enable logging on each platform
• Monitor invocation counts and errors
• Set up alerts for failures

Cleanup

AWS

aws lambda delete-function --function-name signalwire-hello-world

Google Cloud

gcloud functions delete signalwire-hello-world --region=us-central1 --gen2 --quiet

Azure

az group delete --name signalwire-hello-world-rg --yes

Requirements

• Python 3.11+
• Docker (for building deployment packages)
• signalwire-agents >= 1.0.10
• Platform CLI tools (see CLI Setup above)