Soil moisture monitoring for watering plants using MS01 and THR316

I am testing and now using a soil moisture sensor (MS01) with temp/humidity switch (THR316) to inform me when to water one indoor plant. It works up to a point but a lot could be done to make it user friendly and more useful.
I originally thought I could use it in the ground outside to turn on an irrigation system controlled by the ZigBee Smart Water Valve, of which I have two. They work well (with some reservations) so I am loathe to fully automate until the soil sensor system and the SWV are improved.
The main problem is that the THR316 seems to be designed for the THS01 sensor which monitors air temp and humidity. Soil “humidity” (moisture) range is markedly different to typical air values.
Not only that but the soil sensor reports only humidity, not temperature. And the reported soil “humidity” even when absolutely soaked immediately after heavy watering maxes out at <50% which is ridiculous. The typical range is 15% dry (need to water) to 19-20% wet (a few hours after watering) - rather narrow.
The calibration option in the app does not work properly for the soil sensor. Calibration for soil moisture should modify the upper (wet soil) and lower (dry soil) set points thereby widening the displayed range between upper and lower limits.
The app display currently reports only the last 24 hours, so useful trend data and rate of soil drying is not available. This would be very useful indeed so that decisions to water or not, and when, can be made more effectively. It would be useful if the app reported soil moisture with a range of day, week, month, perhaps even more, say 6 months?
Here is another suggestion. Perhaps design a distribution unit for multiple soil sensors selected with a mechanical switch. Or offer a version of the THR316 with up to three inputs. If all goes well for my indoor plant(s), I may make a selector box using RJ10 sockets for multiple MS01 sensors.

Interesting to see that even in the app you couldn’t obtain a correct calibration. I tried with this sensor (MS01) connecting to a ESP8266 using Tasmota, which in theory should allow me to set these limits, but the results were not useable. Makes me think that this sensor is not one of Sonoff best products.

I think you have missed my point. I am sorry I did not explain the problems I have noted so far adequately. The soil sensor itself seems to work reliably and consistently in my limited experience. I presume the soil sensor is a variable resistance depending on soil moisture level.
The major underlying issues for useability (and broad user satisfaction → increase sales) in the real world is how the software reports the input data from the soil sensor. (I do not have an air sensor.) The difference in numeric value (and the graphically displayed) between “too DRY” and “sufficiently MOIST” for a plant is too narrow. And the reported value is not “humidity” but “moisture”.
I am suggesting that the range of displayed numbers needs to be expanded for the soil moisture sensor in comparison to that for an air humidity sensor. The current calibration routine modifies an offset, not a multiplier. Thus there really needs to be a selector in the app for the two types of sensor leading to two paths in the THR316 software. Or it needs to automatically respond to the absence of temperature signal from a connected sensor.
But more than all that, the graphical display and history MUST span more than 24 hours. It needs to be at least two weeks which is about how long it may take for the soil to dry out in my home. A month would be better!

My observation was related only to the sensor itself, like any off the shelf moisture sensor, connected to a Tasmota board.

Exactly, but as a linear analog measurement in Tasmota I found it unusable, but that doesn’t mean it can’t work. As you correctly say, and using Sonoff’ 's TH316 ( originally was the TH10) as you do, what matters is how the software reports the input data from the soil sensor.
Hence my surprise that your results were not up to expectations.