Gladys Assistant
A privacy-first, open-source home assistant
Start new thread

Gladys Assistant 5 - Private, self-hosted smart home, no YAML required

by
•
Gladys Assistant is a free, open-source smart home platform that runs on your own mini-PC or Raspberry Pi. Zigbee, Z-Wave, Matter and 90+ community integrations, a phone-first dashboard, no YAML, and it keeps working when the internet goes down.

Add a comment

Replies

Best
Maker
📌
Hey Product Hunt 👋 I'm Pierre-Gilles, and I've been building Gladys Assistant since 2013, when it was a side project on a Raspberry Pi in my bedroom. The idea hasn't changed: your home should be run by a machine you own, in your house, that keeps working when the internet goes down or when a company changes its plans. Gladys Assistant 5 is our first major version in almost six years: 🏠 A new interface designed for the phone in your pocket and the tablet on your wall 🧩 92 community integrations. Anyone can build one, in any language, and publish it on GitHub 📡 Zigbee, Z-Wave and Matter built in, plus Hue, Shelly, Sonos, Kasa, Reolink and more 🧠 Scenes with triggers, conditions and if/then/else, from a visual editor, no code The software is free and open-source. An optional subscription, Gladys Plus, adds encrypted remote access, Alexa and Google Home, backups and an AI assistant. It's how the project is funded: no investors, no ads. Most of our users are in Europe today, so I'd love to hear from you: what would you need to run your home on Gladys?

Hey Pierre-Gilles, I have partially migrated my HA to Gladys today.
Very much easier to use than HA, I love it.
I have even added an integration via GitHub today using "".
Keep up the good work.

Maker

 Amazing!! Which one did you add?

 plsoftware/gladys-bhyve

twelve years on the same project, from a Raspberry Pi in a bedroom to a real v5, is its own answer to whether the local-first pitch is sustainable. one thing that stands out: the free tier is fully local and offline-capable, but Gladys Plus adds "an AI assistant" as a paid feature. does that assistant run against a cloud LLM, and if so how do you square that with the rest of the pitch being "runs on hardware you own, keeps working when the internet goes down"?

Maker

Great question! Currently it runs on a cloud LLM, hosted on private compute with no data retention and no training on your data. It's an optional layer though: if the internet goes down, only the AI stops, your devices, scenes and automations keep running locally.

And we're working on letting external integrations provide the LLM, which will make fully local AI possible!