Integrates with Alexa+ to provide a shared household assistant for managing chores, schedules, pickup rotations, grocery lists, and spoken morning briefings through MCP tools.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@HomeBase Who should pick up Kabir at 5:30 from football?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A self-hosted MCP server (Streamable HTTP, protocol 2026-07-28, newer than the 2025-11-25 minimum) that gives Alexa+ a shared brain for the household: fair chores, a schedule that catches clashes, pickup duty that rotates, a grocery list, and a spoken morning briefing. It ships with a simulated Alexa+ smart display (voice in, voice out) so you can see and hear it work.
Track: Alexa+ Mini challenge: Open Source (MIT)
Families don't need another calendar. They need the decisions made: who takes the 5:30 pickup, whose turn
it is for the heavy chore, and "wait, Arjun is double-booked". HomeBase returns a why with every decision so
Alexa+ can say it out loud ("Maya is free and has done the fewest pickups").
Related MCP server: Aura+
Area | Tools |
Family |
|
Chores |
|
Schedule |
|
Grocery |
|
Voice |
|
Plus the resource homebase://summary. Anticipated failures (unknown family member, bad time) are raised as
MCP ToolErrors with a readable message, so Alexa+ can tell the user what to fix.
Open http://127.0.0.1:8000/ while the server runs.
Timeline with one lane per person; double-bookings show as hatched overlaps, assigned rides as dashed bars.
Talk to it with the mic (Chrome or Edge), type, or tap a suggestion. Toggle "Speak replies" to hear answers.
Every reply shows which MCP tool it called. The page's /api/tool endpoint dispatches into the same
MCPServer instance that Alexa+ would reach at /mcp, so the simulator exercises the real tools.
The sentence-to-tool step is a small deterministic parser (web/intents.js), so demos behave the same every take.
Try: "What's the plan today?", "Assign the 5:30 pickup for Kabir from football", "Add a chore: vacuum the stairs", "Schedule school meeting for Arjun at 4 pm", "Add eggs to the grocery list", "Who has the most chores?"
pip install -r requirements.txt
python -m homebase.server --port 8000 # MCP at http://127.0.0.1:8000/mcp, UI at http://127.0.0.1:8000/
python scripts/demo_client.py # second terminal: scripted MCP-client demo over Streamable HTTP
python -m pytest -q tests # Python tests
node web/intents.test.js # parser tests (needs Node, optional)Data lives in a local SQLite file (HOMEBASE_DB to relocate). --stdio is also supported for local MCP clients.
"Reset demo" in the UI (or saying "reset") restores the sample family for clean retakes.
Expose the server over HTTPS (for example cloudflared tunnel --url http://localhost:8000, or deploy to any
container host) and register the /mcp URL as your MCP integration in the Alexa+ developer preview.
homebase/core.py: pure logic and SQLite, unit-tested
homebase/server.py: MCP tools (MCPServer from the official mcp SDK) plus the web routes
web/index.html, web/intents.js: the smart-display simulator
scripts/demo_client.py: end-to-end MCP client over Streamable HTTP
tests/: pytest suite
Built during the Amazon Developer Hackathon window (Oct 2026). All code in this repo was written for it.
MIT
This server cannot be deployed
Family meal planning run by your agent: weekly dinners, household votes, grocery list minus pantry.
Household-aware cooking brain: pantry, meal suggestions, dietary safety, recipes, shopping lists.
AI life manager: tasks, home, health, wealth, childcare, pets & more — on your own data.
The first official family-planner MCP: tasks, calendar, meals, shopping and reminders.
