Buying Supported Robots

The recommended source to buy a supported robot is wherever it is cheapest. In general, when buying something, you should always use price comparison websites, shopping communities, check Aliexpress etc.

Deciding on a Robot

This was last updated 2025-12-31.

One thing that tends to happen relatively often is that people think “I will buy this cheaper model to try this out” and then they suddenly have two robots, because it was good but they wanted more fancy features. So generally, I’d recommend just buying the right robot right away.

You likely will want a robot with a full dock that can auto-empty + wash and dry the mops. For budget reasons or other constraints, other purchasing options are possible, but this is the state of the art and you want state of the art.

But what is the right robot?

That is a question that can only be answered when knowing your budget, availability in your area and probably also willingness to follow the slightly more advanced rooting instructions some models need.

From a pure historical perspective, the zoomed-out progression of “core” models for Valetudo has been:

1. Roborock S5
2. Dreame L10/Z10
3. Dreame L10s Ultra
4. Dreame X/L 40 Ultra/Master

Future possibly Eureka J15 Max Ultra?

Beyond these, on the path there were also other supported models, but these ones were the core models, most users flocked to. If you were just looking to buy a robot that works with Valetudo, picking what everyone else picks was/is the most straightforward choice.

What about the other supported robots?

They of course also exist, but they were either intermediate models soon to be superseded, a bit unsuited for widespread adoption (e.g. due to difficult roots) or just dead-end paths where it is questionable why the vendor even released the model.

Of course, if they are on the supported robots list, you can just buy them, install Valetudo and be happy. There is little to worry about there. It might make sense to do so in your market or with your budget.

Also, if you’re buying used, more supported robots is always better, as that gives a lot more options.

Generally speaking, the robots from one Vendor run essentially the same software; just in different versions and iterations. That means that if you e.g. get a Dreame D10s Pro, it will be very similar during actual operation to a Dreame L10s Ultra, albeit a bit more dated as it received less firmware updates.

Do you have comments regarding individual robots?

Yes!
Here’s a table with some thoughts, ordered somewhat chronologically

Models	Thoughts
Roborock V1	Exceptionally cheap (used), easy to root, absolutely no reason to buy it in 2025 due to its lack of persistent maps
Roborock S5	Cheap (used), easy to root, guess it’s fine to buy today if you’re a person that proudly drives a 20 year old station wagon
Dreame F9, Xiaomi 1C	I guess these exist. vSLAM is vastly inferior to LIDAR-based mapping, but they might fit under your furniture better
Roborock S6,S4, etc	If you have one, guess they’re rootable. Otherwise, not worth bothering with unless you happen to stumble upon one
All the CRL200S	They’re fine. Supply of spare parts is great. Rooting is easy. Slightly worse than a Roborock S5. Same station wagon audience.
Dreame D9 (Pro)	They exist. Guess you can buy them and then have some old-ish robot that cleans.
Dreame L10/Z10	A technological leap forward at the time. Tend to fail due to delamination of components on the PCB. Replacement PCBs aren’t rootable.
Roborock S5 Max, S7	Quite involved rooting procedure. Were quite popular at the time, so decent used availability and parts supplies
Dreame L10s Ultra	Overall good bot, still modern-ish. Dock freshwater supply tends to break if you don’t use it. We’re seeing elevated battery failure rates after 2y
Dreame D10s Pro/Plus	There is not much love from Dreame for these. They exist. No need to buy them new
Dreame L10s Pro Ultra Heat	A stopgap bot. Sits somewhere between L10s Ultra and L40/X40
Dreame L40/X40	Newest Dreames supported by Valetudo. Best features and quality sw/hw you can get right now
Eureka J12	Very cheap as a factory new bot. Easy-ish root. Worth considering when on a budget. Comparable to a Dreame W10 with added Auto-Empty and somewhat worse sw
Eureka J15 (Pro) Ultra	Easy root. Alright if competitively priced. Software could still use polish
Eureka J15 Max	Easy root. Alright if competitively priced. Has all the features on paper. Software could still use polish

Robots I didn’t mention in this table are probably irrelevant but also probably fine?

Okay, but what should I buy?

Probably a Dreame X40 or L40.
If you really do absolutely not want to go the PCB rooting route required for all Dreames, get the Eureka J15 Max instead.
If you’re on a budget, check for offers for the L10s Ultra or J15 Pro Ultra respectively.
If that doesn’t yield anything, buy anything else from the supported robots list.

As said, robots by one Vendor run more or less the same software and consist of more or less the same parts, so just look for an equivalent supported model that has the features you’d like to have. It’s fine. They’re all so very similar. Don’t worry about it too much.

Anything else?

One unfortunate truth about them being all so very similar is that if you want to buy something that will last for a decade or so, your options are a nullpointer. They will probably mostly work I guess for the most part, but there is no option where I would confidently say “Just buy this and you’re set”.

Specifically, replacement parts availability is the key limiting factor. Roborock has been mostly alright there, but getting a replacement Dreame battery that isn’t even more broken from the factory than the one it is replacing has proven to be difficult. This is of course even worse for the W10 (Pro) that has an oddly-shaped battery pack. Good luck finding replacements for that.

The same can be said about oddly shaped consumables, though, for the most part, they’re fortunately often interchangeable. Also, at least those you can probably manufacture at home using a 3d printer if you really try.

There is definitely a lot of potential for hackers and/or businesspeople to create a non-shit aftermarket. Think for example a replacement battery that is actually a holder that allows a user to insert individual 18650 cells without the need of welding.
And of course simple 3d models for brushes and idk how we could build the right filters and dustbags. You tell me.

Anyway.
Some food for thought.

Final meta remarks

At the core of making any (buying) decision there is risk and an attempt to minimize it. Unfortunately, risk can never be fully eliminated; you can just change probabilities and, if you throw infinite amounts of resources on it, make things infinitely improbable.

This of course is not possible for normal people, nor is it necessary. There are very easy ways that get you 95% of the way. Unfortunately, we humans often alternate between “I will not take any risk” and “nothing matters, yolo”, whereas “I will look out for low-hanging fruit and build my systems to gracefully handle failures” does not get the same screen time.

What this would mean in this instance is:

1. Not buying a robot with known fatal flaws that completely destroy the machine - a low-hanging fruit
2. Not buying a robot that exceeds your budget and where you wouldn’t be comfortable to buy another one - the system

Generally speaking, quantifying risks is how you get them managed. You need to assess their scope, their impact and their cost and then use that to assess which mitigations exist.

These could be for example knowing that after 2 years, you might need to spend another 80€ for a new battery and budget accordingly. No need to avoid a robot altogether, because of that “oh but it might break aaah what can I do” feeling.

Affiliate links

If you want to support the project, consider donating using GitHub sponsors instead of using affiliate links as they are part of the advertising industry, which some people might describe as bad.

If Amazon is something that you like to use, here are some affiliate links for it.
The DE ones end up supporting me, Hypfer, while the non-DE ones fund Dennis’ work. Apparently it’s easy for a US Amazon affiliate account to generate links on all the other Amazons but not the other way round.

• Dreame X40 Master DE
• Dreame X40 Ultra DE US
• Dreame L40 Ultra DE US
• Dreame L10S Pro Ultra Heat DE US
• Eureka J15 Max Ultra DE
• Eureka J15 Pro Ultra DE Listing 1 DE Listing 2
• Eureka J15 Ultra DE
• Eureka J12 Ultra DE Listing 1 DE Listing 2
• Eureka E20 Evo Plus DE
• Eureka E20 Plus DE Listing 1 DE Listing 2