Inspiration
Inspiration
Smart home devices often operate in silos—your doorbell doesn't know you're in a deep study session, and your TV doesn't automatically adapt to your wearable's productivity context. We built Alexa+ Sentinel to demonstrate how Alexa+ and the Model Context Protocol (MCP) can unify the living room (Fire TV), the front door (Ring), and personal wearable data (Bee) into one proactive agent.
What it does
Alexa+ Sentinel is an interactive Alexa+ simulated experience backed by a Streamable HTTP MCP architecture (Spec 2025-11-25). With a single natural language intent, Alexa+ executes multi-step JSON-RPC tool calls:
- Deep Study & Focus Mode: Syncs pending tasks from Bee Wearable context, sets the Ring Doorbell to automatically instruct couriers to leave packages at the door, and switches Fire TV to a muted ambient timer.
- Smart Delivery Guard: Verifies front-door package deliveries via Ring and routes a live Picture-in-Picture feed directly to Fire TV.
- Perimeter Night Sweep: Arms Ring security, powers down Fire TV, and compiles a morning briefing from wearable notes.
How we built it
- MCP Server (
mcp_server.py&mcp_config.json): Implemented using PythonFastMCPconfigured withstreamable-httptransport and stateless JSON responses. - Simulated Alexa+ Web Console (
index.html): Built with HTML5, CSS3, JavaScript, and the Web Speech Synthesis API to provide real-time voice feedback alongside a live JSON-RPC 2.0 MCP packet inspector.
Challenges we ran into
Designing a seamless fallback mechanism so that judges and users can test the full MCP JSON-RPC 2.0 request/response lifecycle directly in the browser even without spinning up local ports, while maintaining strict compliance with the 2025-11-25 Streamable HTTP MCP specification.
What we learned
We learned how Model Context Protocol (MCP) simplifies multi-device orchestration by turning disparate device APIs into standardized, composable tools that an LLM agent like Alexa+ can invoke deterministically.
Log in or sign up for Devpost to join the conversation.