Introduction
Welcome to Wrist Agent - a seamless Apple Watch to AWS Bedrock integration system that enables one-tap voice capture, intelligent processing through Claude Haiku 4.5, and automatic creation of Notes, Reminders, or Calendar events.
What is Wrist Agent?
Wrist Agent solves the limitations of Apple Watch voice-to-text capture by providing a comprehensive flow:
Watch Dictation → Shortcut → Lambda → Bedrock → Structured Output → iOS Apps
Key Features
- 🎙️ One-Tap Voice Capture: Direct from Apple Watch with complication support
- 🧠 AI Processing: Claude Haiku 4.5 with optional extended thinking (up to 65K thinking tokens)
- 📱 Native Integration: Seamlessly creates Notes, Reminders, and Calendar events
- 🔒 Secure: Header-based authentication with SSM Parameter Store
- 💰 Cost-Optimized: Lambda Function URLs instead of API Gateway
- ⚡ Fast: ARM64 Lambda with sub-second response times
- 🚀 CI/CD Ready: GitHub Actions with OIDC deployment
Architecture Overview
AWS Infrastructure
- AWS Lambda: Go 1.22+ runtime with custom runtime (provided.al2)
- AWS Bedrock: Claude Haiku 4.5 integration with Messages API
- Lambda Function URLs: Direct HTTPS endpoints with CORS support
- SSM Parameter Store: Secure token management (Standard tier)
- IAM Roles: Least privilege access control
Apple Integration
- Apple Shortcuts: Custom shortcut with Watch complication
- Voice Recognition: Native iOS dictation with error handling
- App Integration: Direct output to Notes, Reminders, Calendar
Security Model
- Authentication: Shared token stored in SSM Parameter Store
- Transport Security: HTTPS with CORS configuration
- Access Control: Function URL public access with token validation
- No Long-lived Credentials: OIDC integration for deployments
Quick Start
Get started in 3 steps:
- Setup Infrastructure - Deploy AWS resources with CDK
- Configure Security - Set up authentication tokens
- Install Shortcut - Add to Apple Watch
Use Cases
📝 Note Taking
Convert voice recordings into well-formatted notes with titles and tags.
⏰ Reminders
Create time-based reminders with automatic date/time extraction.