Home Assistant vs Gladys Assistant

Gladys is genuinely simple to install. It runs with a single Docker command (a `docker run` line or a `docker-compose` file you copy, paste and launch), and the documentation walks you through every step with screenshots and installation videos.

And because Gladys is just a container, you stay free to use your mini-PC, NAS or Raspberry Pi for other things too.

Gladys has a clean, intuitive interface. The whole philosophy is to think about the user before the technical implementation: you never have to dig into logs or edit a file on disk. Everything happens with the mouse.

You can build as many dashboards as you want, one per room or by theme (energy, security, and so on), with no configuration files because there simply aren't any.

Gladys deliberately keeps things focused: you arrange the widgets you actually need instead of wading through endless options. And since it's open-source, anything missing can be added by the community.

Gladys' native integrations focus on what matters: open standards like Zigbee, Matter and MQTT, each one carefully built and tested end to end, designed like a polished consumer product rather than by developers for developers.

Through Zigbee and Matter alone, that already means thousands of compatible devices. The bet is that open standards, Matter especially, will dominate, so Gladys invests heavily where the market is heading, not in closed, cloud-only ecosystems that lock you in.

And for everything else, there are external integrations. This is the biggest change in the project's history: anyone can now write and publish a Gladys integration on GitHub, without a pull request, a review, or my approval. They appear in the catalog inside every Gladys instance, next to the native ones, and the community published more of them in a few days than we had built in the core in six years.

For you as a user, nothing changes: you browse the catalog in Gladys, you click Install, and Gladys pulls the integration, starts it and generates its whole interface (device list, discovery, configuration form). Then you start, stop, update it or read its logs from the same place. No command line, no YAML, no technical skills required.

Freedom on one side, Gladys' rigor on the other: each external integration runs in an isolated sandbox (limited RAM and CPU, read-only filesystem, isolated network). If one crashes, it crashes alone, and it can never take your instance down with it. That's exactly what makes it safe to open the door to everyone.

And if the integration you need doesn't exist yet, you no longer have to wait for anyone: you clone the official template, adapt it to your device, publish it, and it lands in the catalog of every Gladys instance within the hour.

Everything lives in a single "Scenes" tab. A scene can be a manual sequence of actions you trigger from your dashboard, or a full automation with triggers, conditions and actions.

The visual editor is intuitive: you start with a trigger (optional), then chain as many conditions and actions as you want, and you can mix them freely, with a condition leading to actions, then more conditions. It's surprisingly powerful while staying easy to read.

And if you ever need to go further, you can offload complex logic to Node-RED and connect it to Gladys over MQTT or HTTP.

Gladys is a project that genuinely listens. There's an active community on the forum, and a roadmap shaped by real user feedback rather than by what's easiest to build.

Best of all, you get direct support from the founder. If you have a question, even a deep one about the code, you can email me, post on the forum, or reach out on social media, and you'll get a personal answer. That kind of access is rare.

Gladys is 100% free and open-source at its core, forever. There's an optional Gladys Plus subscription for remote access, encrypted automated backups, voice assistants (Google Home, Alexa), AI and Enedis energy data.

The business model is really the subscription, with no investors, no ads and no data resale.