What it is
Separate from Jarvis Engine. They share a name, not code. Jarvis Engine is the local operations platform; this is a voice skill.
Say “Alexa, open my jarvis,” and what you say next goes to Claude. The answer comes back as speech.
It’s also a project-management exercise for me: scope, a roadmap, and shipping in small working pieces.
How it works
Voice → Alexa skill ("my jarvis")
↓ ClaudeIntent (AMAZON.SearchQuery slot)
Alexa-hosted Lambda (Node.js 16)
↓
Claude API
↓
Spoken response
Working now
- Voice in, Claude response out
- Session memory, so it remembers the conversation while the skill is open
- Multi-turn conversation
- A catch-all ClaudeIntent using
AMAZON.SearchQuery, so it isn’t limited to pre-defined phrases
Planned
- A personality of its own: a nerdy wise guy
- Persistent memory between sessions
- Weather, reminders, and texts
Still to add here
An architecture diagram, a demo video, and the roadmap.