That’s what I did at my sister’s house with the smart Wi-Fi dimmer switch my brother-in-law had bought. Once I figured out the orientation of the circuit, I could swap that in for the line-side 3-way and leave the “dumb” 4-way and other 3-way switches in place. It’s nice how that works, dimmer function and all.
More later, but for now…
This is so cool! Commence time sink!
It’s a nice day so I’m headed outdoors for awhile, but I’ll post back later tonight. So far, I’ve successfully integrated my Wyze devices (except Sense devices) and my Google Home nest speaker and nest thermostat. It all just works! I will play around later with automations, turn off IFTTT and see if I can get things working with HA geofencing alone.
I’ve looked around at integrating Wyze Sense v2 devices and coming up short. It seems like the old v1 USB hub is needed. I found some obscure posts on GitHub from around 2021, but they never seemed to be released for public use.
I may also start looking into flashing Wyze devices with custom firmware if I can get them integrated into HA. I like Wyze hardware, but I’m ready to end my reliance on the cloud. Internet outages have occurred often enough that I just want these things to work as long as they have power.
It does sound like a lot of your frustration is stemming from the unique situation in your house, and not necessarily from the relays themselves (or, at the very least, that the relays didn’t solve your unique challenges in a way you had hoped for).
Since my last update, I’ve done a lot of reading and YouTubing. As far as I can tell, Shelly relays allow an old “dumb” switch to be converted into a smart one, and at a much lower price point than brand new switches. While this does seem compelling (and there is apparently a huge group of people that swear by the relays - see “best smart switch” discussions everywhere), I’m personally unconvinced because of two reasons:
- Converting an old switch means keeping the old switch. I have a house full of the most basic on-off switches possible, and an upgraded aesthetic would be nice.
- The functionality will be strained: it is possible for a switch to physically be in the “on” position but be turned off. If I then wanted to mechanically turn the switch back on, I’d have to turn the switch off and on again, or change the functionality so that every flip of the switch simply toggles the state, regardless of physical orientation.
I know these are tradeoffs to save a bit of money. For me, if I’m only replacing a handful of switches, it seems more practical to just get the right thing. So unless I’m missing something here, I’ll be going the direction of replacing switches altogether.
Seeing as how you’ve already purchased them, I feel like it’s worth trying to find a solution. Although, the piece about resetting the relay after a power outage is a deal breaker for me. As far as figuring out the house’s wiring, it may be worth spending some time trying to get it to work with regular dumb switches so you can understand what’s going on, then try to add the relays back in. Or, use the relays for the basics (non-3-ways, for instance) and spend a bit extra to get the full switch in the more complicated spots.
It seems like these are the two forerunners. Both are expensive - $50-$75 per switch. Lutron requires the use of a hub and is more widely available and supported. Inovelli requires (recommends? offers?) the use of a bypass device, installed at the light fixture, to “help with switch to bulb performance” in non-neutral setups. The programmable LED on the switch is a cool feature, but I’m not sure I need it. I wish there was a simple Zigbee switch that looked like a classic rocker style switch and had the reliability of Lutron, at a lower price point. So for now, I need to do a bit more research on the options.
I was able to get everything going within HA. It seems the global Notifications On/Off toggle within the Wyze Integration is broken, as it has no effect within the Wyze app. Toggling notifications for an individual camera does work, so I should be able to set automations to do that for each camera anytime I leave or return home.
Does this work for the second generation of sensors? I don’t have a USB receiver, only the newer Sense Hub
I’m ok with the sensors not working within the Wyze app, as my eventual goal is to discontinue daily use of the app and move everything into HA. If I can’t accomplish that with the Wyze sensors, then I’ll end up getting new sensors. As it is, the sensors can’t really be used to automate anything, and that was the whole point in having them.
I imagine this doesn’t work the same way with Google Assistant? I don’t use Alexa at all and don’t plan to.
I explored HA automations a bit, and was able to figure out basic leave/return automations. For testing purposes, I had HA send a push notification to my phone when the automation ran, and it seems very responsive in just two days I’ve been testing it. Today, I’m going to disable my IFTTT automations and transfer the logic over to HA. This is, after all, the solution to my original question!
@carverofchoice A follow-up question regarding RTSP. I remember you saying that RTSP is a little quirky with Wyze and can tend to mess things up. I had gotten it working within HA while trying to figure out device integration and just left it. Meanwhile, one of my flood cams started acting more erratic than normal. I was often having to hard-reset the device to get the light to come on at all. I ended up disabling Docker Wyze Bridge, and things have been more stable again, but I can’t figure out if the two were connected.
I also vaguely remember seeing you say that you don’t use a Wyze subscription anymore. Correct me if I’m wrong on this. For me, the only reason I still have a subscription is so I can get live notifications to my phone. Generally, I check to see that I recognize the activity and then dismiss the notification. I don’t need cloud storage, nor do I need Wyze’s apparently-fancy but practically-useless AI detection features. Each camera is recording to its own SD card, which is sufficient in the event that I want to review footage of something happening.
So my question is: how could I move toward completely replacing my Wyze subscription with HA? All I need it to do is monitor cameras for motion and send a push notification when detected. Bonus if I can employ some sort of logic to ignore recognized vehicles, people and pets. I also need a stable way to view the live feed of my cameras, both locally and remotely (this is in fact the only thing that I like about Wyze and the Wyze app - and think I’ve seen you say that this is admittedly something that HA struggles with a bit). Is there a way to do this without losing the ability to open the cams within the Wyze app for that one single purpose?
I specifically want local control and access of my devices. To think that I can’t even view the SD card footage without an internet connection unless I physically remove the card from the camera is silly.
I already have remote access setup through DuckDNS working.
The only significant issue I can think of is that RTSP causes a camera to live-stream constantly This can be taxing on cameras’ limited resources, particularly when it is also used in other ways (if it’s still doing cloud streaming plus other live streaming to another source like your phone sometimes, etc) but generally should be okay as long as the camera is receiving sufficient power (ie: not connected to a long third party USB cord that gives it insufficient power to do multiple things at once).
Considering this is a floodlight that should be getting enough power, I’m guessing that’s not the issue, but that’s what I generally find to be the issue if they’re related.
That is incorrect. I have Cam Unlimited, Cam Plus Lite, Home Monitoring Service, and Sprinkler Plus. But I also use RTSP.
I suggest considering having your RTSP feeds go into a detection integration. One of the most popular ones on HA is Frigate. Another one is Scrypted. Both of those will allow you to do your own customized detections and such for improved notifications. However, if all you care about is motion notifications, you should be able to do this with just your cameras integrated into HA through Docker Wyze Bridge. You would just set the trigger to be the “STATE” when that camera entity’s “motion” state changes to “on” and then have the action be a service call to the mobile app’s notify. ie: something like:
alias: Wyze Cam Motion Detected
trigger:
- platform: state
entity_id: binary_sensor.wyze_cam_motion
to: 'on'
action:
- service: notify.mobile_app_my_phone
data:
message: "Motion detected on Wyze Cam!"
title: "Wyze Cam Alert"
The difficulty is getting your phone to receive notifications when you are away from home, but you said you already have that resolved to DuckDNS, so it should work for you.
You would still be able to use the Wyze app to view your SD card, but you won’t be able to access the SD card any other way. I suggest you use something like Frigate or Scrypted or another method to start saving recordings of your important cameras to a computer harddrive or NAS so you can access the recordings through Home Assistant/Frigate/Scrypted instead of having to rely on the SD cards.
Welcome to the club. So worth it though. It’s really exciting overall. You won’t regret it. But yeah, huge time sink.
The Wyze sense integration DOES work with the second generation of sensors, however, not with the second generation hub. You have to connect them to the old Wyze sense bridge that the v1 sensors used…you plug that into Home Assistant and then connect the sensors to it. BUT I have to say that this integration is not actively maintained, and one of the Home Assistant Updates earlier this year broke it, so you have to manually update some code in the integration or use one of the forks.
This is a big reason why I want to push Wyze into an official integration with Home Assistant at some point. It definitely won’t be on their roadmap before second half of next year at the soonest though, but if they did an official integration then we could have direct local access to a lot of stuff that wouldn’t have to “poll” their server. But alas, they won’t even consider anything until at least CSA’s Matter 1.2 launches. So we’ll see.
In general, I would suggest trying a workaround instead or getting new sensors. If you go the new sensor route…if you aren’t allergic to buying things from China, I would suggest considering looking at aliexpress.us for some of your smart home needs. Otherwise you can buy the exact same things on Amazon from resellers/drop-shippers for an exponentially higher price and more quickly…but I buy the exact same things from Aqara for way cheaper on Aliexpress, or other generic smart home things using zigbee, etc. Another good one to consider is Ikea. They have some great Zigbee sensor too. If you buy in bulk (I think $50 minimum?) they will ship them to you instead of make you go in-store.
This may get fixed in the next update. I know there is one scheduled that is fixing a lot of toggle bugs/issues, and we have one of our Wyze Volunteer Mavens to thank for those fixes coming up soon.
Sounds like you’re coming a long nicely. There are so many awesome things you’ll start doing.
Wow. That was a brutally long post.
Two bad they don’t havr page numbering or percentage bars for message read.
I almost fell asleep. Guess I will go to bed early.
lol
I got asked a bunch of questions and probably didn’t even answer them all, and not very thoroughly either. I thought I was being to short on my responses to him.
Well, @carverofchoice is a Maven for good reason. He has helped me many times right from the beginning when I joined these forums, and he’s thorough and gives good feedback and advice. That’s rare these days, and I sure do appreciate it! This thread has saved me a ton of time and stumbling around in the unknown, which was always my main hesitancy to get into HA.
So, many thanks, good sir! This thread could become the basis of a how-to for any Wyze user trying to get integrated into HA without replacing all their hardware.
I will have to continue playing with this, then. The cameras are all on the stock USB cord, plugged into mains power.
To confirm, is Docker Bridge pulling the RTSP stream locally, before it goes up to the cloud? As it is, some cameras are at the extreme limit of my WiFi coverage (though the Flood Cam is not), and my cable upload speed isn’t so hot. If the stream data is going up to Wyze servers and back down to my HA instance, that could be causing a bottleneck.
My goal here is to keep the cam data local as much as possible. I would be OK using the Wyze app to review SD card data for the time being, but I suppose I will have to look into local recording. This is something I’m not quite ready for; I live in a quiet neighborhood, and I review footage very rarely (and generally only for fun).
I will look into this next. Messing around in the Wyze app yesterday, I noticed that the Friendly Face Detection feature has been activated. If this works within reasonable expectations, that would prompt me to keep the monthly subscription. I know it’s new and will need time to improve, but I’m hoping it’s better than the other AI detections.
“Motion Detection” is a component of the Integration, outside of Docker Bridge. I may have to play with that a bit too. The problem really comes down to replacing Wyze’s recording/marking of specific events to eliminate having to seek through the entire SD card. I have to think about that one a bit more and figure out if it’s worth the $10/month subscription.
Well, unfortunately I don’t have the sense bridge and probably couldn’t get my hands on it easily. I would be willing to do some manual code updates, but it’s probably more work than it’s worth. I’ll start looking around for replacement sensors. A workaround may be to use HA to capture the push notification as you described earlier, and react appropriately. I’m mainly interested in getting a push notification from the door sensors only if the person is not recognized. That would be far more useful than knowing how many times the door was opened by someone who is supposed to be there.
For onlookers and to complete this thread, I have finally accomplished what I was initially looking for: I now I have HA toggling Wyze Cam notifications based on geofencing, using the HA mobile app on my phone. It’s not quite as responsive as Google Home, but more so than IFTTT was (and free). I also now have Google Home integrated into HA, and I’m sure I can find a way to trigger a HA script using the Google geofencing, and then there would only be one app on my phone that is polling my location.
@Crease By the way, I ended up getting an outdoor Wyze plug, which came yesterday. I’m planning on playing around with it today. I’m actually adding some outdoor light to the chicken coop area, and figured that would be a good excuse to get one.
@carverofchoice Thanks again for all your help! Going forward I will probably start a new thread for any new questions regarding Wyze/HA integration, or will start posting on HA’s forums or using the Discord server to figure out HA issues.
I second that. He pointed me over here after I asked a question about Video Doorbell v2’s integration with Google Home in Wyze’s Discord server, and then I did a write-up to contribute back to a topic I’d used a year prior when deciding to buy and install Light Strips (those are still working well) because I like the idea sharing and giving-back spirit in this community.
I just noticed that Home Assistant has been added to this topic’s title.
I like those, and I still have a 3D-printer project in mind with those that I’d like to sit down and design someday. The original Plug was my gateway drug into the Wyze ecosystem; I’ve been using those since 2019.
You’re the third person I’ve read here in the past month or so who’s mentioned a chicken coop. My limited experience with those is going out to one to hear some music.
Thanks for continuing to share your updates! I enjoy reading about how you’re accomplishing your goals and making this stuff work for you, and now that you’ve mentioned integrating Google Home into Home Assistant, I’m even more curious about the latter.
YES! The stream is all local. The only caveat here is that the initial authentication when the camera boots up still goes through the Wyze servers to approve the camera to have permission to stream. After that, some people report being able to shut off the internet from the cameras and they will continue to stream locally through RTSP. But they do need that initial internet connection when they reboot just to authenticate they are approved to function. So, the constant video stream will not use up your internet bandwidth, at least as long as the device it is connecting to is ALSO on the same network without using device isolation, etc.
Wyze and every other camera company would go broke if their video streams had to go through their server and internet bandwidth. No, it’s peer to peer.
One thing to keep in mind with this is that if you have both Person detection AND face detection notifications turned on with the same camera, then the camera can detect a person from farther away (before it can make out a face yet), and it will send a person notification, then when you get closer it will detect the face, but since it already sent out a notification for this event (when it detected the person), it won’t send the face event (even though it will still detect, label and categorize the face detection for the event, just not send the notification if the person notification already went out seconds earlier). WyzeMatt said they are going to resolve this in an upcoming update so people don’t feel like they have to remove person notifications just so they can get reliable face notifications, but if face detection is important to you, then you should know about that. Face Detection is optimized at about 300+ pixels, which they said on a 1080p camera is roughly 2 meters away for the average face. It may detect some farther away than that, but at a lower confidence interval.
This would be a good reason to use something like Frigate integrated into Home Assistant. Then you can set up an automation that checks if a recognized face was detected on that camera within the previous X seconds. That actually sounds like a great automation. I might go update some of my Swimming Pool automations to only alert me if an unknown face is recognized in the pool area, or tell it not to alert me as long as my face or my wife’s face were recently detected in the area too. Then it will not alert me when we’re there, but will tell me if any of our kids or other people are near the pool without our supervision. I should probably go do this soon. I even have a contact sensor on the pool gate, so I could pretty much set up the same thing you are thinking of.
You could also do that with Wyze notifications of a face using Macrodroid an a Webhook (URL) that triggers a Home Assistant automation. For example, maybe I would create a helper binary toggle that is something like “carverofchoice on Pool Cam” and have it toggle on when there has been a face detection for me within the last 5 minutes or so, and automatically toggle off after 5 minutes of no new notification with my face on that camera. Then a separate automation can check if my face has been detected on that camera in the last 5 minutes (the helper toggle is on) and if so, then it doesn’t notify, but if it is off, then it will notify me that someone else opened the gate. Something like that. That’s a good idea to reduce notifications I don’t want and only get more accurate ones. Thanks for the idea.
That’s awesome! I’m glad you we found a reasonable solution for you and that I was able to help you justify making the jump:
I’m definitely a n00b compared to the help some of the home assistant experts can give, but I knew I could at least give you enough to get you started on the Wyze side of things. I’m glad we found something that will meet your needs while being able to still use the stuff you already have.
Thanks for calling me in here to help @Crease
Sorry for my delay in responding! I’ve been picking up extra shifts so I can justify buying some Home Automation gear.
For any onlookers, I will say that if you like Wyze hardware but want increased control, HA is the way to go. I absolutely should have made this jump sooner, and it basically allows you to do anything you can think of.
This is good to know. I already have the Wyze Cam docker installed, but I couldn’t quite tell if it was fully working. I need to spend more time playing around with it. My next step would then be to get Frigate working. While reading and watching videos, it’s apparent that a lot of people use and love Frigate.
I started using Wyze’s new Friendly Face feature, and provided some sample footage of family members arriving at the house. I will admit that Wyze made this very straightforward, and I’m actually impressed. After setting it up so that I don’t get notified when a friendly face is detected, it seems as if I’m getting far fewer notifications when family comes over, though I haven’t confirmed this by looking at my events log. However, the activity in the backyard still triggers a slew of notifications, so the goal of the HA automation would be to just mute those notifications until everyone leaves (I need to ponder how to tell HA that the friendly guests have left and to resume normal notifications).
If there is something specific you want to do using a Google Home/Home Assistant integration, let me know what it is. I can try it out and report back whether it’s doable and how hard it is.
The biggest challenge with integrating Google Home and HA is that Google needs an externally available link to get to your HA server. So, that means you have to either use Nabu Casa’s cloud service ($6.50/mo or $65/year) or create a subdomain that is updated dynamically as your public IP changes (free). I went with the latter, and while it took me an hour or two to get it all figured out, you can definitely do it. My biggest hurdle was trying to figure out what I needed to do, not so much how to do it. I ended up jumping into the HA Discord channel and asking for help for the last bit. Someone immediately responded and got me the rest of the way there (after installing an add-on, I had forgotten to start it.
Once I had that figured out, the process of integrating Google Home into Home Assistant was straightforward, with the help of the very good documentation. You do have to gain access to the Google API using an API key, creating a project, etc., but if you can follow directions and take things one step at a time, you can do it.
I am extremely impressed with how polished Home Assistant is at this point. Documentation is very clear, not only for the things that are built-in, but also for integrations and add-ons. The HA dev community has definitely made a conscious decision to make this project accessible to less tech-savvy people. So far, I’ve been able to do a natural language Google search for whatever it is I want to do, note the integrations and add-ons that will accomplish that, and go over to Home Assistant, find said integration/add-on, and read its corresponding documentation. Most of the time that has been enough, and if it still left me a bit unsure, I searched for a YouTube video on “how to configure xxx in Home Assistant” and found many excellent results every time.
I figured out my first real automation last week. I now have a light that turns on based on time of sunset and turns off based on the length of day combined with the time of sunset. This would have been very difficult or impossible, but I was able to do it in Home Assistant very simply.
Details, if you're interested
I bought a Wyze Outdoor Plug and installed it at the chicken coop. Attached to one outlet are lights for both the front of the coop and in the covered portion of the run. I am using a Wyze Switch in smart device mode, installed at my back door, to control this outlet. I can now come home after dark, hit that switch, and go do the chores without a headlamp (collect eggs, fill feeder/waterer). Attached to the other outlet is a light inside the coop. There is where HA came into play.
As daylight decreases, egg production is reduced and may stop altogether. While somewhat controversial, one can add artificial light at a daylight frequency to keep egg production going year-round (especially in more moderate climates). I decided to try this, because as much as I love keeping chickens, I do it for egg production, not to have pets. It’s a financial decision for me.
Typically, folks use a timer to control how long the light will stay on after dark. The challenge to this is that the timer needs to be adjusted often, maybe even weekly. The goal is to have a total of about 14-16 hours of combined natural and artificial daylight each day, and as the number of hours of daylight decreases, the length of time for the artificial light needs to increase. The time that it turns on also needs to be adjusted each day, based on the changing time of sunset.
I accomplished this simply with the following automations:
- Turn light on 1 hour before sunset (the coop is in a shady area and gets dark long before sundown)
- Turn light off 10 hours before the next sunrise - which would be roughly 14 hours after the sun rose.
This is definitely just the beginning. I’m doing a lot of research into locally-controlled sensors and switches. I got my Zigbee hub, although I haven’t been able to try it out since I don’t yet have any Zigbee devices.
Not to hijack your topic from its original subject or anything, but since you offered…
I was reading recently that there’s a Feedreader integration for Home Assistant, so one thing that would push me toward trying it out now is the idea that I could potentially have Home Assistant consume RSS feeds of schedules for my favorite sports teams (college basketball is about to start!) and then automatically change some interior LED colors (bulbs, light strips) to team colors shortly prior to game start times. I don’t know if that’s doable, but I also don’t know why it wouldn’t be, and it seems like a fun project to try at some point.
I already have Routines in Google Home set up for this and have for some time (currently disabled for off-season), but in the past what I’ve done is wait until after a game has started (like during a timeout or commercial break or something) or has finished and then manually updated the Routine’s start time for the next event on a team’s calendar. Automating this from an iCal or RSS feed would be cool.
What I suspect would be possible is just triggering the lights from Home Assistant with no need to involve Google Home at all, but, again, I’m just speculating from zero experience with this.
That’s a cool use case for your chicken coop lighting. I enjoyed reading that. Thanks for sharing!
I’m going to do some looking around and I’ll let you know. I have everything I need to try this myself. I’m with you - I don’t see why it couldn’t be done.
How would you filter down all the games in one day? I.e. are interested in one team, one league, one region… I really think that’s probably the biggest hurdle - looking at only one game at a time.
I also want to add in my Google Nest Mini smart speaker. It’s be cool if the automation included an announcement through the speaker and a push notification… “It’s game time!”
Edit: take a look at this post from the HA forums.
I wouldn’t…or, at least, I don’t think I would. Not knowing much at all about Home Assistant, I would expect that it could be configured to poll an RSS feed of a team’s schedule or conference schedule on a daily basis (or however often a user configured the polling) and then parse the XML based on the poll date and time and create/delete a scripted lighting event that would be scheduled to run prior to game time. (I’m thinking deletion could be for past games and/or schedule changes.) I don’t know anything about Home Assistant scripting, though. I’m just making guesses about how I think I’d approach the problem.
I do this with my Google Home Routines. The last Action in one of them, listed after the lighting changes, does this:
Maybe I should see if I could get it to play the fight song, too.
I actually read recently in another topic that it’s possible for Home Assistant to generate notifications, so that seems like something that could be built into an automation.
Thanks for the link. I’m not ready to dive in yet, but I skimmed that Reddit post.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.