Voice automations + profiling and configuration

since this is my first post, ill have multiple questions, so please bare with me.
I am trying to achieve a smart home unit, by which I can control a relay, read IR remote controls, and then transmit them, using IR . my questions.

  • I want to initialize a “virtual remote”, in other words, when a user wants to store any kind of remote, “AC remote”, he shall just initialize it first, so that the MCU can store them like a dataset of some sort, resets, then displays this remote UI, I want this to be of course stored, “initial idea would be to store it on flash from MCU side” but I have 2 problems :

    • If I needed to erase it from flash, how can I do so on esp32 without doing a mass erase ? i’ve encountered this problem with wifi provisioning which I can’t find a solution for as well.

    • I have no clue whether this idea is applicable UI wise or not, if yes, any help with how I can start ? If you can provide me with an example for a button let’s say “power on, off button”. can you provide me with a coding example ?

    • do you recommend any pre-configured libraries on ESP32 that I can add on vscode by which I can have some defaults “supported remotes”, unless if nothing works, the user then customizes his own remote however he wants. at the end this is a bunch of IR codes. I am mainly looking for ACs and TVs, nothing big.

    • how can I achieve this UI wise on mobile app ? could that be achieved ?

  • I cannot seem to find voice automations, and i cant find anything about alexa or google voice assistant.
    I tried to take a look at the upgrade plans, but I can find nothing for developers, or pros, white label has nothing other than “contact sales”. and it wont make sense to go for white label since am not an enterprise, its just a simple thing for my home. I want to use the voice assistants in general (alexa, siri, google voice assistant) without the use of IFTTT, I am trying to generalize my software as much as possible. so that I can make it generic. and easy to use.

Your questions seem to conflict in some ways.
On the one hand you seem to be describing a generic product which could allow the user to choose from preset functions as well as learning new IR codes, but then on the other hand you say that this is for personal use, and that you want to generate your code yourself.

These two comments seem to conflict slightly, because on the one hand you seem to want to build something that has a fully functional UI (which makes it sound like it’s going to be a commercial product) but on the other hand say that it’s just for personal use.
Whilst it may be nice to build a UI that does everything you’ve described, I would have thought that in the first instance you’d want to focus on functionality rather than elegance?

Also, you refer to voice activation and talk about Alexa, Siri and Google. Are you actually using all three of these for voice activation at present? Most people tend to settle on one system, otherwise things can get quite messy.

Voice integration isn’t available in Blynk unless you subscribe to the Enterprise plan, in which case it’s an add-on.

If this is for personal use then the solution you choose will probably depend on whether you have other home automation projects in mind.

Personally, I use Node-Red and Alexa, along with Blynk, to achieve a very similar solution to the one you describe.
The Blynk and Alexa plug-ins for Node-Red allow excellent integration, and Node-Red acts as a “rules engine” your home automation system.

I use IR transmitters that are based on Wemos D1 Mini’s and use these to control devices line ceiling fans, AC systems etc. I could extend that to TV and entertainment system control if I wanted to, but personally that’s not something I’m interested in at this stage.
My IR learning tends to be done offline, with a Wemos D1 Mini used to analyse the codes from unknown devices, then these are manually entered into the rules engine.
Most devices don’t require this approach though, because the IRremote library already has built-in support for most protocols.


Hi Pete,
while yes it might sound confusing to you. But I need to first do it for my own home, and yes, I dont want to keep on coding, it shall be wired into my house electricity. In addition to that. I am an embedded systems engineer myself. So it’s kinda what I do for a living. and I always learned how to do it rather generic to be highly customizable and doable. I hope that explains to you the Why’s

as of to why I want to use the 3 voice activations. that would be for me and my wife, as we have different phones with different OS. that shall explain the next point.

Now for the point of the UI. I dont want to use something more than blynk. but I want to know how I can achieve it using blynk’s UI. for example, the wifi provisioning has its own page.
can this be achieved using blynk? if yes how ? that was my question.

as for Node-Red, I dont have much experience with it. but ill give it a shot. thanks for that.

As for the ir transmitters and protocols. I think ill dig deep myself with the libraries there. I just wanted a quick sketch/C file.

like IRremoteESP8266. it looks good. but I cant find a good example. thats all :slight_smile:

I don’t think you’re going to achieve this this using Blynk.
Not because of the limitations of Blynk, but because of the way that the IRremote library works.


Then I guess I’ll write mine :wink:

You should take a look at Node-Red.
If you have a machine you can run it on then it simplifies the process enormously, because all you need to do is send an MQTT message to your IR sender telling it what type of IR message to send. Your IR sender is then a pretty dumb device, and the logic sits in Node-Red’s rules engine.

You the get all of your voice assistant integrations and you can use them to trigger the flows that send the appropriate MQTT messages to the IR sender. You can use Blynk as your UI if you wish, or do that another way.

I still have my doubts about how practical it will be using three different voice assistants. They all work differently, and getting consistent logic from all three is going to be hard work!