Inspiration Bridging the gap between voice-first platforms and open standards by leveraging the Model Context Protocol (MCP) to turn standard smart speakers into context-aware, agentic collaborators.
What it does Amazzz connects the Alexa+ ecosystem directly to custom backend tools and data sources via a self-hosted MCP server using Streamable HTTP, allowing users to execute complex multi-modal actions effortlessly through voice.
How we built it
Implemented a compliant Streamable HTTP MCP server capable of exposing dynamic tools and data resources.
Integrated the Alexa+ MCP toolkit to bridge the assistant's client-side orchestration layer with our custom server endpoints.
Handled multi-modal rendering states to ensure smooth card and widget updates directly on screen-enabled companion devices.
Challenges we ran into
Adapting standard SSE-based mental models to the newer Streamable HTTP transport specification required careful state management.
Ensuring low-latency round trips for voice interactions while processing complex agentic payloads.
Accomplishments that we're proud of
Successfully standing up a fully compliant Alexa+ MCP add-on from scratch.
Achieving seamless real-time execution of backend tasks triggered purely by natural voice prompts.
What we learned
Designing for voice requires hyper-concise payload designs since conversational responses demand immediate, low-token clarity compared to text-based UIs.
Transport protocols like Streamable HTTP dramatically simplify session handling for cloud-hosted agent architectures.
What's next for Amazzz
Expanding the tool registry to support advanced proactive agents that can trigger workflows autonomously.
Integrating rich MCP App widgets for immersive visual dashboards on Alexa Show devices.
Log in or sign up for Devpost to join the conversation.