HA-Jev
Turn Jev answers into Home Assistant sensors and automation results.
What Jev decides
A yes/no probability, one option, or a score for a supplied question.
HA-Jev exposes probability, choice, and score questions as sensors, automation actions, an AI Task entity, and an Assist conversation agent.
- Jev role
- Answer configured questions about home state for Home Assistant automations.
- Input
- A configured question and relevant Home Assistant state or automation context.
- Output
- Typed Noul, Choice, or Score result used by a sensor or action.
- Decision frequency
- When a configured question, sensor update, or automation action invokes Jev.
- Fallback
- Automation behavior and thresholds remain in Home Assistant configuration.
How it works
Overview
HA-Jev integrates TypeSafe Jev with Home Assistant. A configured question can become a sensor whose state is a probability, selected option, or score. Automations can call Jev actions and use the typed response; an AI Task entity and Assist conversation agent expose the same model through Home Assistant interfaces.
Jev makes the narrow judgment. Home Assistant owns the resulting automation, and users configure the question and any threshold. The integration reports API usage and an estimated daily cost against a token budget.
- Language
- Python
- License
- MIT
- Author
- AboveColin
Sources
Related projects
Kill My Idea
Turn a startup pitch into a compact set of scores and a blunt Kill, Fix, or Ship verdict.
What Jev decides
Dimension scores, idea category, whether the pitch is understandable, and supporting diagnostic signals.
Jev Chat Assistant
An Android conversation copilot that uses Jev to interpret chats and rank reply candidates without sending for the user.
What Jev decides
The likely intent, conversation risk, reply need, best next action, unresolved tension, and which of three candidate replies fits best.
Distill
Route coding-agent work, reasoning effort, and context retention through bounded Jev decisions.
What Jev decides
Which configured model and effort should handle a call, what evidence deserves attention, whether another attempt is useful, and what context should be retained.