Talk to Core Lattice — a simulated Alexa+ experience

What it does

Talk to Core Lattice is a working simulated Alexa+ experience in a web app. You talk to Core Lattice — a proprietary conversational AI architecture I designed and built myself — and it talks back: it routes what you say to the right capability, remembers your name and your session, walks you through a guided demo, builds structured plans from a goal ("plan my hackathon week"), explains its own architecture, and — by design — answers honestly when it can't verify something instead of inventing a confident-sounding answer, which is more than can be said for some landlords.

How it works

The official rules for this hackathon expressly allow a simulated Alexa+ experience built with your own tools, exempt from the Alexa runtime-technology-hook requirement. I took that path, and the complete simulation source is in my repository: a conversational engine (intent router, session memory, scripted demo flows), a zero-dependency Python server exposing a chat API, and an Alexa-style web interface with a status ring, suggestion chips, and voice in/out through the browser's Web Speech API. What judges see in the video is exactly what the repo runs — no mockups.

The significant update during the submission window

Core Lattice itself predates this hackathon — it's the AI architecture I build and use in my own work. What is new, and built entirely inside the October 2026 submission window, is the Alexa layer around it: the simulated Alexa+ experience (interface, voice layer, conversation design), plus the hackathon scaffold of the engine — the intent router, memory, and demo flows packaged so anyone can run and talk to the system in a browser. That packaging is the submission: an existing proprietary engine, made conversational through Alexa+ for the first time, during this hackathon.

Why it matters

Most conversational AI demos are a thin front end on someone else's model. This one is the opposite: an independent developer's own architecture, built for her own real workload — including litigating her own civil case as a pro se plaintiff (I couldn't afford a lawyer, so I built one) — opened up so anyone can talk to it the way they'd talk to Alexa. It's a working proof that the next conversational interfaces don't only come from big labs.

Built With

• core-lattice-?-the-author's-proprietary-conversational-ai-architecture-(hackathon-scaffold-in-repo)
• javascript-/-html-/-css-(vanilla)
• muse-(meta-ai-assistant)-?-pair-programming-debugging-and-documentation
• python-3-(standard-library-only-?-no-external-dependencies)
• web-speech-api-(browser-speech-recognition-+-synthesis)