← Back
T.A.R.S
Active Development Role: Creator
Personal AI assistant handling ambient automation across my life.
Local LLMsPythonHome AssistantFastAPI
14:02:11 [SYS] Idle
14:15:00 [CRON] Syncing tasks
14:30:45 [NET] Checking pings
[EVENT] 3:45 PM · Calendar: Meeting with Advisor
The Problem
I wanted a “second brain” that didn’t just store information, but actively worked for me. Existing voice assistants are limited to basic commands and lack context about my ongoing projects, communications, and daily routines.
What it Does
T.A.R.S is a continuously running background service on my home server. It listens for ambient triggers—calendar events, emails, sensor data—and takes proactive action using local LLM inference.
- Contextual Awareness · Summarizes threads and drafts prep notes before meetings.
- Ambient Automation · Controls physical environment (lights, displays) based on focus state.
- Private by Default · Runs locally on my apartment server without sending sensitive personal data to external APIs.
Architecture
(Architecture diagram placeholder)
[T.A.R.S Architecture]
Outcome
In active development and used daily. It has reduced my manual task-switching overhead and acts as a true automated assistant.