TeslaFi vs TeslaMate

[kcseay]

I've used TeslaFi for almost 2 years. I've been using Teslamate in docker for about 2 months. I am impressed with teslamate over some of the other self-hosted attempts I've tried. It is more stable and hasn't crashed on me.

For TeslaFi, I'm paying for subscription for 1 vehicle. I love the graphics and the ease of getting your information. But you have to pay for each vehicle. In Teslamate, I'm continuing to learn as I set up and configure things. I have it integrated with Home Assistant and have just configured with my phone to disable falling asleep if it detects my bluetooth getting into the car. I use it on my multiple Teslas. I would hope to stop the subscription for TeslaFi, but until I can figure out how to do everything in Teslamate that I like in TesalFi, I'll continue to pay, but only for 1 vehicle. The developer has actually incorporated a request of mine of a feature that I use in TeslaFi.

I wish I was smart enough to develop an app like that on my own. As far as privacy, I feel that if you're not using the original Tesla App, you're taking the same additional risk with any of theses apps or companies.

 

[Darmie]

Totally this. If anyone thinks they can do a similar or better job for $50 of your time, please go for it. For me that's worth about 15-20 minutes of time. But if they do, they're all still using the same unlicensed Tesla API and subject to all the risks and challenges that entails.

I also don't understand all this agita over the "unlicensed" Tesla API. There are dozens of mobile apps and web apps that all access the API, and probably hundreds of other developers who have rolled their own API programs. If it stops working, they all stop working. So what? You're going to cry over $50 sent to Teslafi for his exceptional work if it suddently stops working? At this point, if Tesla closed all public access to the API, there would be a huge groundswell from thousands of people -- not just the app developers, but the thousands of affected users. I don't think at this point, that's in Tesla's best interest, even if they never intended to have 3rd parties access the API.

I second that. Been using them for over 2 years on two different vehicles. Always trusted and true with no fancy candy.

 

[aegidius]

it suddently stops working? At this point, if Tesla closed all public access to the API, there would be a huge groundswell from thousands of people

Not worried about Tesla closing the API so much, but more that they will change some aspect of it without warning that breaks things. It is, after all, an undocumented, unsupported API and therefore subject to change. They have not undertaken to keep compatibility or issue versions, etc. to outsiders.

 

[HankLloydRight]

Not worried about Tesla closing the API so much, but more that they will change some aspect of it without warning that breaks things. It is, after all, an undocumented, unsupported API and therefore subject to change. They have not undertaken to keep compatibility or issue versions, etc. to outsiders.

Exactly. But the same risks exist for anyone using a third-party app or is doing it themselves. So what's the brewhaha over not using third party apps because Tesla doesn't support or document the API??

I sent the TeslaFi support an email today asking if they have licensed access of the Tesla API. Unless they can prove to me they are legit users, they won't see me renew past the trial.

I'd be happy running my own thing on the understanding it might stop working.

 

[raynewman]

Exactly. But the same risks exist for anyone using a third-party app or is doing it themselves. So what's the brewhaha over not using third party apps because Tesla doesn't support or document the API??

It does depend how many people you want to give keys to your precious.

 

[HighlyCharged]

Exactly. But the same risks exist for anyone using a third-party app or is doing it themselves. So what's the brewhaha over not using third party apps because Tesla doesn't support or document the API??

There's two points that everyone is missing.

1. Teslafi is a paid service. They obviously do their best and have previously worked around unannounced changes to the Tesla API, but there is the risk that the API can change enough, or be taken down, such that it stops working. Some people have stated that they don't care about losing money in that case, which is fine. It's also fine if you do care.

2. Privacy concerns. Teslafi is hosted in the USA and is not subject to the same privacy provisions as other countries, and is subject to USA law which is very different from where I live. The privacy statement on their web site says that they will only share data to comply with the law, provide your services, protect your rights (whatever the hell that means), and to fulfil business obligations. This is one of the loosest privacy statements I've ever seen. Again, you may or may not care about this. Some people do.

The self-hosted solutions, like TeslaMate, are open source software and your data is your own. Obviously you have the hassle of setting up, which is fine for those who have the smarts to do so. Given that you don't pay anything and that your data is self-hosted, you literally have nothing to lose by using it.

So, you pay your money and you make your choice. Everyone has an opinion on what works for their own situations.

 

[HankLloydRight]

There's two points that everyone is missing.

First, I'm not missing those points.

Given that you don't pay anything and that your data is self-hosted, you literally have nothing to lose by using it.

Second, the point you seem to be missing is that the self-hosted options are not without cost. They might not have a monetary cost, but they have a significant time cost. For most people, the monetary cost for Teslafi (and others) is significantly less than the time investment needed to set up and maintain the self-hosted solutions. And if the API goes down, all that time and effort is lost -- no different than the paid offerings. So you are far, far from "literally have nothing to lose by using it".

They obviously do their best and have previously worked around unannounced changes to the Tesla API, but there is the risk that the API can change enough, or be taken down, such that it stops working.

How is that any different than with a self-hosted solution? If the API changes (which it will) the people who offer a paid solution have a powerful incentive to fix or address it as soon as possible. With the self-hosted solutions, you're pretty much out there twisting in the wind waiting for someone to volunteer to fix the open source code and then you have to update or re-install, more time and effort invested and a possible data outage waiting for someone to post a patch or fix. There have been plenty of open-source projects that have died on the vine like that when the original contributor abandoned the project and there wasn't anyone willing to pick it up.

 

[Flatbat]

Anyone is/has running both of these and have any opinions on them?

I have neither of them, so no opinion. So six weeks on @bigbones are you still using them both? Ignoring the security concerns for a minute, can someone give me brief rundown on how the two products compare. Pros and Cons etc. Features that make one better than the other. I only have but a single M3 so won't need to look at multiple cars.

 

[Vostok]

According to Electrek, Tesla is considering opening its own App Store, which of course would necessitate Tesla formalising and documenting the API, creating an SDK, have stable releases, etc.

5 things Tesla is bringing to market in 2020 - Electrek

 

[HankLloydRight]

According to Electrek, Tesla is considering opening its own App Store, which of course would necessitate Tesla formalising and documenting the API, creating an SDK, have stable releases, etc.

5 things Tesla is bringing to market in 2020 - Electrek

Just to spell this out as it actually is -- TESLA isn't considering or has announced anything. There aren't even any rumors or such.

This is all based of an off-hand comment Elon made during the RideTheLightning interview:

In order for it to be worthwhile for somebody to write an app, there has to be enough of an install base to warrant the effort. Even if you are going to port something, it’s still got to be worth the effort. As our number of vehicles grow, it starts to potentially makes more sense to develop games and other applications for Tesla. We just need a lot of cars.

So well before Tesla even considers an API/SDK/App Store, they have to reach some magical unknown number of "a lot of cars." That's not going to happen in 2020, and certainly not something Tesla is "bringing to market in 2020".

That's not even including all his weasel words "starts to potentially make more sense."

Even if Elon Musk tweeted right now "Yes, we'll have an SDK and App Store launching in 2020" that means it won't happen for at least two years after that, if at all. "3 months maybe, 6 months definitely… "

They can't even get FSD software out the door after how many years? And that's their best hardware and software teams working on that.

I shudder to think that the firmware/UI team will be responsible for the API/SDK/App Store.

 

[HighlyCharged]

I have neither of them, so no opinion. So six weeks on @bigbones are you still using them both? Ignoring the security concerns for a minute, can someone give me brief rundown on how the two products compare. Pros and Cons etc. Features that make one better than the other. I only have but a single M3 so won't need to look at multiple cars.

My initial impressions of both are

• Teslafi has a much better UI experience. The data presented from TeslaMate is via Grafana dashboards on a separate URL. If you've never used Grafana it's quite jarring (I have, and it's still jarring!).
• However, the Grafana graphs are a lot prettier and more informative in their own right
• Teslafi does costing for charges and drives. TeslaMate does not (although it's on the roadmap to add)
• TeslaMate was a real pain in the arse to set up if you want to access it outside of your house - but this was mostly due to my internal DNS setup and the crappy documentation
• TeslaMate records a ton of data, but not all of it is presented to the user.
• TeslaMate can start polling after a wake-up a lot more quickly than TelsaFi, because you can set a trigger from your phone when it connects to the car's bluetooth. TelsaFi has missed the start of drives, that's not happened with TeslaMate (so far).

Being open source, anyone can add stuff to TeslaMate so it's the sort of project that could really take off if it gets popular. However, the dude who's written it has chosen Elixir as the coding language, which is right up there with the best of the most obscure languages you could possibly pick. I would have contributed by now if it was in say, Python, but hardly anyone seems to know Elixir. This is a serious negative point for me.

So in summary, TeslaFi is better as things stand, but you need to pay for it. TeslaMate is free and has good potential, but you need some technical smarts to set it up.

 

[Flatbat]

Thanks @bigbones Nice to have finally have a post comparing the two products, rather than two pages of banter about APIs.

One the basis of your mini review I have signed up for the Teslafi trial. about to jump in the car and drive 350kmh as soon as the battery is topped off. So for a quick setup of Telsafi I have:
- Australianised the units to metric.
- Checked the box in Settings > Sleep modes to Teslafi Sleep mode Enabled.
- Added a home address
I gather I need to turn off cabin overheat protection in cars onboard settings too. For now I don't need anything being scheduled from Teslafi as my schedule is too erratic.

Anything else I need do to get up and running quick. PS Has anyone got a spare referral code thye could PM me?

 

[HighlyCharged]

I gather I need to turn off cabin overheat protection in cars onboard settings too

Yep that's all I did.

Will PM you a referral code in a sec.

 

[qdolan]

I got TeslaMate up and running a month or so ago but after a while it started spewing timeout issues and stopped working. I attempted to look at the code to trace the issue but didn't feel like learning another language so I just turned the docker containers off and will look at it again another time.

 

[Flatbat]

Thanks fellas for the referral codes. Turns out (much like Telsa) I actually needed the code prior to signing up (Stupid me) which makes sense now I think of it. Did my 315km drive across Victoria and it is all nicely logged in Teslafi reporting 100% accuracy.

 

[Flatbat]

After two days with Telsafi, the thing that is going to bug me is the way it creates so many drive events. I gather that I can combine two drives in to one but no more. Is this correct? So I went for a drive into town 50km away. Got there and made 3 stops quick stops, then back home = 5 drives!! I guess I just have to put up with this. Is combining tow drives reversible in Teslafi?

 

[jmmmmm]

Im now using TeslaMate and trying to get a better understanding of the data. My driving efficiency seems to be poor and I'm not quite sure what to do to change that. Any ideas on how to improve my numbers or are they common.

 

[Vostok]

Im now using TeslaMate and trying to get a better understanding of the data.

Does TeslaMate understand the metric system of units and measures?

My driving efficiency seems to be poor and I'm not quite sure what to do to change that. Any ideas on how to improve my numbers or are they common.

From what I’ve read, that “efficiency” percentage in the table is not “efficiency” at all. It is the actual distance travelled (“mi”) divided by the reduction in estimated range on the battery indicator (“TR Loss”). So if you drive 100 km and the range on the battery indicator decreases from 400 km to 300 km then it would say your “efficiency” is 100%. I don’t know what one would call that calculation, but “efficiency” is absolutely the wrong term for it.

In an EV “efficiency” is measured in Wh/km. I can’t see that in the table anywhere.

In my first road trip this weekend, I covered a total of 680 km at an efficiency of 150 Wh/km. I consider that exceptionally good, we travelled at highway/freeway speeds and with the AC on most of the trip. The reduction in range reported by the battery indicator (summed up over individual segments) was about 5% more than the distance we travelled. I guess that would give a TeslaMate “efficiency” of 95%.

Not using the AC improves efficiency a bit. One leg we did without the AC on (because we didn’t need it) was 147 Wh/km whereas a leg with it on (and probably driving a bit faster) was 164 Wh/km. Driving faster definitely hits efficiency. Driving at 110 uses noticeably more energy than driving at 100.

The other thing to try is to turn sentry off when parked for extended periods in safe locations. Sentry is an absolute power hog, and you will lose kWh while travelling no distance. I don’t know how TeslaMate deals with power consumed while parked, whether it ignores it or whether it gets rolled into the “efficiency” of the next trip segment, which would kill it.

 

[Flatbat]

Im now using TeslaMate and trying to get a better understanding of the data. My driving efficiency seems to be poor and I'm not quite sure what to do to change that. Any ideas on how to improve my numbers or are they common.

To improve your numbers move to a warmer climate
It is literally freezing in Norfolk and cold is bad for efficiency. The sweet spot for me seems to be in the 30 to 35 C / 85 to 95 F band while it is way colder this where you are in Northern hemisphere winter. Your drives distance are also very short so your battery will rarely be getting up to an optimum temperature and will be continually wasting energy in trying to do so.

Teslafi has a chart which shows this by comparing the outside temp to rated efficiency and also displays Wh/km.

 

[Anubis]

Interesting... here is mine... I get good efficiency between 15 - 20C and 30 - 35C.