Deconz 3 wishlist/thoughts

Firstly, many thanks for your continued efforts on deconz. For me, it’s the best way to manage ZigBee devices. Wide range of supported devices, lightweight, nice hue-compatible API.

I saw you were working on a version 3, so I thought I’d float a few ideas that may or may not be helpful.


  • GUI as a web app. I’ve not needed the GUI until recently, but it was interesting to see it. However I feel the accessibility of the GUI is limited due to being part of the server component. If it were fully REST API driven with a web app client, like Phoscon (or maybe part of?), then it would be much easier to use, consume less resources on the “hub” and probably be easier to develop.
  • Ability to have “virtual” scenes support, for devices that don’t play nice with ZigBee scenes.
  • Perhaps storing device definitions in yaml would be a little more concise?
  • Perhaps using a document/no-sql db would allow easier migrations? (I think I’m assuming you’re using sqlite)


  • I find the multiple brands confusing. Dresden Electronic, deCONZ, Phoscon, Conbee. I know these serve difference purposes, but from a consumer point of view, it’s not that clear. Perhaps consolidating onto a single brand name would be helpful? It might afford other opportunities too. Think about how Hue is so cross-cutting (and short and memorable) as a brand.
  • An app - being hue compatible is great as you get support of third party apps more easily. But an own brand app (as well) would perhaps make the story more complete and more accessible.
  • I tried out the Dresden Electronic ZigBee Controller and Scene Switch. I really wanted to like them and give you guys more support in buying them. But both of them weren’t good enough for my usecase.
    • The controller seemed to have a low PWM frequency and not particularly solid connections.
    • The switch had a great physical layout - but the buttons were squishy to press and made multiple clicking noises on press. Also the price was really high given that I would prefer a different style backplate.
  • Since deconz/phoscon are essentially free, you must make your money from the hardware. The conbee dongle is great, but it feels like a missed opportunity to make market leading zigbee accessories.

Thanks for the feedback! Forwarded it to the appropriate channels :wink:

I’ll follow this with interest :slight_smile:

This is a personnal POV.
Desktop applications are more powerfull and fast than web application.
And I can be wrong (I don’t know all possible libs) but I think developping a C++ desktop application is easier than a server destined to a Web application.

Have you already compared Word on desktop and Word online ? It depend of the browser, but it s not the same comfort (for me, just my personnal opinion)
Web application are fine for small application, or have better ergonomic/aesthetic, but on complex application they are limited.
Adding more feature on phoscon yes, but I don’t think it s possible having the same thing than on the GUI.

I’m stepping in here and hide some comments as we are going way off topic.

I too would like to see deCONZ move our of a VNC session into more of a web interface. Even if that does not happen, a feature I’d love to see is just a simple “Submit Device Request” button.

deCONZ already has all the data you need to request a new device, so why go through the screenshot process when the data could be submitted directly with no errors on our part? (I’m looking at myself here. :grinning_face_with_smiling_eyes: )

1 Like