Hi Everyone!

I equipped my whole house with Tapo L630 bulbs + a few Tapo L530. These bulbs are really cool and have some nice features!

Unfortunatly some of the more advanced features like effects are not yet implemented in python-kasa or do not work as described in the API documentation. Since I've spent a considerable amount of time playing around with their features I would like to share my findings with the community.

Custom light effects can be set:

kasa --host xxx --json command edit_dynamic_light_effect_rule '{"id":"L2","color_status_list":[[100,300,100,0],[100,240,100,0],[100,170,100,0]],"change_mode":"bln","change_time":2000,"scene_name":"Relax"}'

This command will modify effect L2 (Relax by default).

The color_status_list is an array of color states the bulb transitions through.
E.g. [100,300,100,0] translates to 100% brightness, 300 hue, 100% saturation, 0 colortemp.

The maximum number of elements in the array seems to be limited to around 10 color states (I'd have to do some more testing to verify the actual maximum number).
Transitions between color temperatures can also be done by setting the last field to a valid color temperature e.g. [50,300,100,3200],[50,300,100,5000] will transition between 3200k and 5000k.

change_mode can be "direct" or "bln". "direct" will just switch to the new state without transition. "bln" will slowly blend to the next state.

change_time is in miliseconds and can be as low as 8ms (!). I think it can also go up to 20seconds+.

scene_name can be any name you like. You also don't have to modify existing effects. e.g. "id":"L3" will just add another effect.

The API itself would support setting custom effects but it does not work right now with L630/530 bulbs. Something in the API returns that the bulb has no custom_effects. Please fix this!

Absolutly crazy effects can be achived with ultra-low change_times and switching between R,G,B color states.
I've tested this in my livingroom with 12 bulbs and 10ms. Everything moving looked like it had some sort of color trail.
Kind of wierd if you see your arms move with R,G,B trails ;-)
(Please be careful though as I don't think this is benefitial to the bulbs health and these ultra-low timed, strobe like effects, can potentially cause epileptic seizure for some people.)

It is also possible to synchronize effects by sending carefully timed commands in parallel to several bulbs or group of bulbs at the same time. I created a shell script that made an effect that looks like a wave of colors is slowly moving through the building. Is there a way to send commands to multiple bulbs by using the python API ?

Default-States (hard power on) can be set with:

command set_device_info '{"default_states":{"type":"custom","state":{"hue":0,"saturation":100,"brightness":100,"color_temp":3300},"re_power_type":"always_on"}}'

The API also would support setting default states (hard, soft) but this does not seem to work with L630/L530 models.
I suspect this is because the bulb does not have a "soft" state, at least I haven't found a way to set it...

Automated mass-provisioning can be done on linux with a small script like this:

nmcli dev wifi list|grep Tapo_ | while read MAC SSID INFO  ; do
 echo "Connecting to ... $SSID (${MAC})"
 nmcli dev wifi connect "$SSID"
 ping -w 1 -c 1 >/dev/null 2>&1 && echo CONNECTED
 kasa -v --host 192.168.0.1 --username XXX --password "YYY" wifi join "My_Wifi_SSID" --keytype wpa2-psk --password "my_wifi_password"
done

Local NTP:
You can make the bulbs use a local NTP server if you dns-mask *.pool.ntp.org and *.nist.gov to a local server

There is an annoying issue with the bulbs caused by their firmware. If they cant connect to the cloud services they become stuck in a reboot loop where they become unavailble for 1 minute every 10 minutes. I tried absolutly everything humanly thinkable to find a fix/hack for this but the only way to improve things slightly was by running a script that reboots all bulbs every 580 seconds. This helps to reduce their "downtime" to 10 seconds instread of 1 minute every 10 minutes.

Comments, feedback, own findings greatly appreciated! Please share what you've got!

Cheers!