shadow carbon
Hi everyone 👋
I’m working on a small open-source Home Assistant project called HA Context Explorer.
The goal is to make a Home Assistant setup easier to understand directly inside Home Assistant and later through structured exports for AI-assisted analysis.
It is meant to help users and AI assistants understand:
- entities
- devices
- areas
- integrations
- relationships
- automation/script references
- logic context
- what is used where
- where changes might have an impact later
The project is designed to be:
- strictly read-only
- local
- privacy-first
- based on authenticated/admin-only access for real HA data
- focused on understanding, not editing or controlling devices
Right now, it exists as a Home Assistant custom integration with its own sidebar/panel UI for direct exploration. A structured export function is still planned, because one important use case is providing useful context to AI tools such as ChatGPT, Claude, or local LLMs without exposing more data than needed.
The current version is still early and experimental, but it already has working views for overview, entities, devices, areas, integrations, relationships, and a first logic/reference slice.
Important:
I’m not a professional developer. This is my first real project, and I’m building it step by step with AI tools while learning.
I’d really appreciate feedback from people who know the Home Assistant ecosystem better.
I’m especially curious about:
- Does this idea make sense for HA?
- Are there existing tools doing something similar?
- Are there architectural pitfalls I should avoid early?
- Are there better ways to expose this metadata safely?
- Is the read-only / privacy-first direction reasonable?
- What should I watch out for when adding exports later?
GitHub:
https://github.com/wdani/ha-context-explorer
Any feedback, ideas, or warnings would be greatly appreciated 🙂
