An "ethical" mixer / swap protocol - private but not helping crime

[d5000]

From discussions in our beloved Meta thread Mixers to be banned (link to first relevant post) I got an idea for a mixer business model which could be both private and almost safe to not support criminals, at least the "big ones". The initial idea was combined with an idea by @joker_josue. Just today I refined the model a bit more based on thoughts about other posts in that thread.

This mixer concept would have 2 goals:

1) Be able to offer high privacy standards: a no-log or almost no-log (i.e. delete logs every day) policy, and no registration nor KYC checks of any kind (with the exception of cases where users don't follow the rules of the service).
2) Be able to block most coins which come from theft and other forms of crime, without having access to the coins themselves. Thus they wouldn't be forced to seize them.

The mixer or swap would have the following rules:

1) Users can mix or swap small amounts without any KYC check. "Small amount" would be really relatively small, I would set the limit at about the equivalent of $1000 per day or less, perhaps only $500.
2) The remaining users who want to mix more than $500 or $1000 per day would have to go through a pre-screening of the UTXOs they want to mix, done on a form on an independent website operated ideally by a different party, where the UTXO is analyzed by a chain analysis algorithm but without a deposit. If the funds are detected as high-risk, i.e. directly coming from criminal sources, they will be rejected. All other coins get a token [1] and they are allowed then to mix.
3) KYC will only enforced if the users deposit coins which break both rules, i.e. they were recognized as high risk, more than $500/$1000 and the user deposits them without having a token.
4) The pre-screening is done via a web form where the address or the UTXO can be entered, or an API. This ensures the mixer has never access to coins which could trigger an obligation to seize them. This means: The final deposit of the coins you mix will happen only if the pre-screening has succeeded or the funds are small enough.

For rule 1 to work, there is an additional rule:

5) Funds that pretend coming from different users in a single day, cannot come from the same origin address, i.e. the users cannot simply try to identify as "different users" and split their coins open. This means that the coins being mixed in a single day by any user (without pre-screening) cannot be from the same origin.

And there is an important additional "rule" to prevent the mixer being silently collecting data:

6) The service should be Tor-only and usable without JavaScript. Above all the pre-screening "service" would only be allowed on a website with an onion address.

Example: Let's say an user has an address with 1 million USD worth of BTC. He moves them around during 72 hours and tries then to open 1000 mixing instances on the mixer. The first mixing process would go through, but from the second one all others will be rejected because the coins come all from the same origin.

There could be an exception for coins which come from old origin addresses, for example if the coins were already split 2 months ago or more. But that would normally not be a problem, as all the algorithm wants to ensure is that the same user only mixes $1000 per day without pre-screening. This would also prevent the situation that somebody has received a payment some time in the past and mixed the change coins just the same day the person who paid the coins wants to mix the change. However, for that to happen one of both users would have to have very bad luck that exactly the same day the other party would have mixed the coins

What d'you think? I think such a mixer could be what many Bitcoin users want: 100% KYC-less for small funds mixed by retailers, and still with a high chance to be relatively "safe" from law enforcement. For criminals, the mixer would be to slow to "pre-launder" the funds through them.

Of course the protocol would not be completely legal in jurisdictions where all custodial Bitcoin services are forced to do KYC, but there are several countries such a service could be still run.


A small detail:

[1] Technically, the token could be simply consist of the UTXO the user provided, signed by the certificate of the mixer or the entity operating the pre-check. This avoids that there is any direct data connection between the pre-screening instance and the mixer itself - because the mixing website would recognize the signature of the pre-screening entity, if the UTXO is the same one than the one pre-screened.

[retaur]

I think you're going to have a problem with the limiting splits part. What if two, four or a lot of users decide to withdraw from a gambling site that's not too well known or that's new and you're there trying to do cryotonalysis on all their transactions or just blacklisting them and confiscating funds.

Maybe the way some non kyc exchanges do it where they ask for kyc or reverse the transaction otherwise (but that might become costly and another reason for users to complain - what if the funds get reversed back to an exchange account and then neither of you have them).

[RGBTC]

Conceptually, it makes sense to only prevent vrime from laundering their "dirty" bitcoins. The next problem:

and still with a high chance to be relatively "safe" from law enforcement

The final score, however, will be determined by AML audits. Afaik, Chainalysis, Elliptic, and other AML companies too often (I think it's become the default) label coins that have been involved with mixers as medium-high risk. To this point, there's nothing you can do unless you don't want to touch centralized services at all. Implementing this protocol doesn't significantly improve the chances of passing an audit until their systems are also adjusted to the new protocol. So, improvements must occur on both sides.

[ABCbits]

2) The remaining users who want to mix more than $500 or $1000 per day would have to go through a pre-screening of the UTXOs they want to mix, done on a form on an independent website operated ideally by a different party, where the UTXO is analyzed by a chain analysis algorithm but without a deposit. If the funds are detected as high-risk, i.e. directly coming from criminal sources, they will be rejected. All other coins get a token [1] and they are allowed then to mix.

Are there any cost or limit of UTXO pre-screening? If it's free, the website/service could receive DDoS by sending tons of pre-screening request.

Of course the protocol would not be completely legal in jurisdictions where all custodial Bitcoin services are forced to do KYC, but there are several countries such a service could be still run.

Do you have list of country where such service could run without being illegal?

[hugeblack]

This system already exists. GingerWallet.io has a "COIN FILTERING" feature with a default coordinator coinjoin, which filters out high-risk transactions from coinjoin transactions. Therefore, the outcome will be better than random coinjoin selection, *provided GingerWallet collaborates with some block analysts.*

Some legal details are here ---> https://github.com/GingerPrivacy/GingerWallet/blob/master/WalletWasabi/Legal/Assets/LegalDocumentsGingerWallet.txt However, I don't know if governments would be interested in this or if they simply don't want any customer privacy.

The question will always remain: how do you ensure that high-risk are included? Is any transaction blacklisted by governments truly high-risk? At some point, the system will be abused, and two values ​ for the same BTC will exist (cleanBTC  dirtyBTC)

[d5000]

I think you're going to have a problem with the limiting splits part. What if two, four or a lot of users decide to withdraw from a gambling site that's not too well known or that's new and you're there trying to do cryotonalysis on all their transactions or just blacklisting them and confiscating funds.

I think you refer to several users withdrawing coins from the same service and then get flagged due to "same origin address"?

There could also be a pre-screening for small coins which apparently , separated from the high risk coin detection, i.e. you could offer another web form where you can first enter the UTXO and check if it doesn't come from the same origin address than already mixed funds. Probably this will be needed, unfortunately, because it requires technical knowledge to rule out the possibility that your coins come from the same origin than others that have already been mixed this day.

And yes, in the case this rule is broken but no AML check has been performed, a direct refund should be the next logical option.

Are there any cost or limit of UTXO pre-screening? If it's free, the website/service could receive DDoS by sending tons of pre-screening request.

Good point, but I think the work needed to prevent this would not be more than what block explorers have to do to prevent being DDOSed.

Do you have list of country where such service could run without being illegal?

No. I guess you could check kycnot.me and see where the services are located. I think where this would be illegal, Bitcoin lobbyist groups could push for legal changes to allow such services, but of course they need to be strong enough. There being a chance for them to be legalized or not criminalized is just the reason why I would set the daily limit quite low, approximately as low as the common limits for cash payments without KYC in shops in the strictes countries (e.g. some in the EU, like France).

@hugeblack: Looks interesting, thanks for mentioning! However this protocol is meant for custodial mixers, which have some advantages, above all for people without the required technical knowledge to perform CoinJoins.

The final score, however, will be determined by AML audits. Afaik, Chainalysis, Elliptic, and other AML companies too often (I think it's become the default) label coins that have been involved with mixers as medium-high risk.

The risk of "pre-laundering" coins from crime in such a service would be low, approximately as low as swap exchanges with "shotgun KYC", so they could be in contact with the AML audit companies to not include this kind of mixer in their "blacklists".

[davis196]

2) The remaining users who want to mix more than $500 or $1000 per day would have to go through a pre-screening of the UTXOs they want to mix, done on a form on an independent website operated ideally by a different party, where the UTXO is analyzed by a chain analysis algorithm but without a deposit. If the funds are detected as high-risk, i.e. directly coming from criminal sources, they will be rejected.

Do you have any idea who is going to run this "independent chain-analysis website"? How much is going to cost to analyze the origin of the coins and will the cost be included in the mixer's fee?
The whole idea of an "ethical mixer" seems weird to me. If the users are going to mix clean coins, then why do they want to mix the coins?
If you want to hide your tracks it seems obvious that you are doing something illegal and the coins are obtained in a rather illegal way. That's the whole point of crypto mixers. I don't see any point in mixing clean coins.
Anyway, good luck with this project(I'm assuming that you really want to build such "ethical mixer").

[NotATether]

So basically it will be like Tornado Cash but with Tether-style blacklisting?

Who will decide the addresses to be blacklisted? A person? A committee? This is a sensitive topic, on top of the matter of not really having any native way of being able to swap Bitcoin for other cryptocurrencies (which would lead to being able to swap "bad bitcoins" for "good bitcoins") except for Thorchain I guess.

[d5000]

So basically it will be like Tornado Cash but with Tether-style blacklisting?

No, it would be a custodial service. And the blacklisting should be done in a way the operators never get the chance to confiscate the coins if the users play by the rules and deposit/mix the coins only after pre-screening or within the daily limits.

Who will decide the addresses to be blacklisted? A person? A committee?

Do you have any idea who is going to run this "independent chain-analysis website"? How much is going to cost to analyze the origin of the coins and will the cost be included in the mixer's fee?

(answering both as you asked basically the same)

The service could be run by a chain analysis company directly, sponsored by mixers who cooperate. And of course the mixer's fee would include all costs including this "sponsorship". Take into account however that many no-KYC swap exchanges do the same thing already now, only in a less transparent way, and they can offer competitive fees.

This is a sensitive topic, on top of the matter of not really having any native way of being able to swap Bitcoin for other cryptocurrencies (which would lead to being able to swap "bad bitcoins" for "good bitcoins") except for Thorchain I guess.

I would consider atomic swaps a "native" option. However, with atomic swaps if you swap to Monero and back for example (or use Thorchain which probably has the same problem) you can't be sure that the coins you get back are really "clean" (until you analyze the origin yourself, of course ...).

The whole idea of an "ethical mixer" seems weird to me. If the users are going to mix clean coins, then why do they want to mix the coins?

Here we disagree fundamentally. All coins being traceable means basically that chain analysis would eventually be able to trace every payment and every user. And blockchains are public. Now even if you are okay with that -- what if a chain analysis' database is hacked and leaked and your name appears connected with some payments you did and you wouldn't like your friends or your boss to see?

Also what if you're living in an authoritarian country (more than 50% of the world population do)? Let's say you supported the opposition or even an human rights organisation with coins -- is that "illegal" for you too then and worthy that you're "doxxed" by chain analysis and then perhaps imprisoned?

In conclusion, without mixers or other privacy techniques, Bitcoin would be an immense surveillance infrastructure with no fungibility and privacy at all. It would be Big Brother's dream. And I'm sure if we go so far, usage would eventually drastically decrease, and the price will crash too. Censorship resistance is the most notable quality or USP of Bitcoin and other decentralized cryptocurrencies, and it would be gone. (The 21 million limit? Altcoins have done that too.)

See: https://bitmixlist.org/mixers-necessity.html from the "Mixers to be banned" thread for more arguments.

BTW, I'm not going to build it. The idea is that those than want to build a mixer with a slight chance to be not inmediately criminalized (and perhaps to be present in this forum again), could take inspiration from this.

[examplens]

What d'you think? I think such a mixer could be what many Bitcoin users want: 100% KYC-less for small funds mixed by retailers, and still with a high chance to be relatively "safe" from law enforcement. For criminals, the mixer would be to slow to "pre-launder" the funds through them.

Some mixers, Jambler partners, have the option of no service fee testing. The amount must be exactly 0.001 BTC. I assume that this can be considered a small transaction, there are no special conditions, so you can make several orders from the same address.

Honestly, I can't see a mixer and any KYC in the same context under any conditions.
The only 'improvement' I would understand is the AML check of the coin before executing the order. So the mixer, if it recognizes coins from illegal streams, would return it back to the sender.

[d5000]

Honestly, I can't see a mixer and any KYC in the same context under any conditions.
The only 'improvement' I would understand is the AML check of the coin before executing the order. So the mixer, if it recognizes coins from illegal streams, would return it back to the sender.

In the protocol I'm proposing, the AML check is done even before the deposit, like an address check on a block explorer (e.g. Blockchair) but for the transaction/UTXOs. Or none at all is done, if the "small amount" rule is respected.

I'm mentioning KYC only because in Bitcoin you can't avoid people depositing coins on an address, so as soon as the mixer publishes a deposit address, the mixer could transfer any funds to it, including not-pre-screened high risk coins. Of course simply returning the funds back would be ideal. But if the mixer is legally required to seize/demand KYC if the coins are "very high risk" (e.g. directly stolen funds) then in this case a minimal KYC infrastructure for these cases would be unavoidable.

I basically am thinking about the same model some no-KYC (or "almost no KYC") swap services are enforcing, but with the "pre screening" possibility before the mixing operation is done.


In theory, one could build a protocol where even this "minimal KYC" is not needed, if the mixer requires a "contract" transaction which gives the user the opportunity to "auto-refund" the funds himself. It is a bit complicated but would not even be user-unfriendly because the "magic" would be hidden to the user:

1) The user provides a refund address to the mixer.
2) The mixer generates a P2(W)SH address with two branches, where the coins can be claimed both by the mixer and by the owner of the refund address.
3) The user deposits coins to this P2WSH address.
4) If the mixer thinks the coins are high-risk, he will NOT move the coins and print out a message to the user. The user can withdraw the funds now at any time, and as soon as he withdraws them, he of course loses the right to mix the coins (so there's no double spend risk and no timelock or so is needed)
5) If the mixer thinks the coins are ok, the mixing operation is carried out, and both sides are happy.

[Lucius]

I remember that there were some discussions on this topic years ago in terms of how to prevent bad guys from using mixers and laundering hacked coins, but I think it all boils down to the fact that the mixers of that time (and today) simply did not even think of going in that direction. There is only one reason, and that is profit.

My humble opinion is that most of the mixer's profits come from problematic transactions and their profitability lies precisely in that source. Of course, there could also be what you call an "ethical mixer", but no matter what measures would be implemented in it, the authorities would hardly accept any way of mixing with the aim of trying to hide the origin of even those coins that are not tainted according to them.

I would honestly like to see someone try to make such a mixer as an experiment to see if something like that could be profitable and not be on the government's radar.

[d5000]

My humble opinion is that most of the mixer's profits come from problematic transactions and their profitability lies precisely in that source.

There were actually studies about that (unfortunately I have no link right now), and it seems that in the early days mixers had only a relatively small share of problematic transactions. But in the late 2010s and above all in the 2020s that share increased up to around 50% in some cases.

The problem is however that also legit privacy-focused users may have left mixers because of the increased problems with "mixer-tainted" coins on major exchanges. I think they now use either CoinJoins or Monero swaps. The protocol proposed here aims to fit into that niche.

Of course, there could also be what you call an "ethical mixer", but no matter what measures would be implemented in it, the authorities would hardly accept any way of mixing with the aim of trying to hide the origin of even those coins that are not tainted according to them.

I remember FIOD (the Netherlands financial crime unit) declaring here in the forum that crypto mixing was not necessarily illegal, and I have seen this stance also in EU resolutions, so it depends on the jurisdiction probably.

It would also depend on the type of taint - "mixer taint" or "CoinJoin taint" ("mid-risk" coins) for example is not necessarily illegal sourcing. Stolen coins are another story.

[Lucius]

~snip~
I remember FIOD (the Netherlands financial crime unit) declaring here in the forum that crypto mixing was not necessarily illegal, and I have seen this stance also in EU resolutions, so it depends on the jurisdiction probably.

It would also depend on the type of taint - "mixer taint" or "CoinJoin taint" ("mid-risk" coins) for example is not necessarily illegal sourcing. Stolen coins are another story.

Has it ever been confirmed that that BTT account belonged to them? It didn't make any sense to me then (and it doesn't even make sense now) that such an agency would create an account on this forum to announce that it had seized a mixer.

Regarding the EU and cryptocurrencies, the truth is that there are laws that everyone should apply - but the application of those laws still depends on which member country you are in. However, allegedly from the beginning of next year, all members would have to be fully aligned with these laws and implement strict AML/KYC with all services related to cryptocurrencies.

Of course, there will always be those states within the EU that do things their own way, but in general, the EU currently has much bigger problems than cryptocurrencies - they are fighting drones that they don't know where they come from, what their goal is, and how to disable them. Three years ago, a large drone flew over three EU/NATO member states and crashed a few hundred meters from a student dormitory - so much for what politicians and defense are like in the EU.

[examplens]

Of course, there could also be what you call an "ethical mixer", but no matter what measures would be implemented in it, the authorities would hardly accept any way of mixing with the aim of trying to hide the origin of even those coins that are not tainted according to them.

I remember FIOD (the Netherlands financial crime unit) declaring here in the forum that crypto mixing was not necessarily illegal, and I have seen this stance also in EU resolutions, so it depends on the jurisdiction probably.

It would also depend on the type of taint - "mixer taint" or "CoinJoin taint" ("mid-risk" coins) for example is not necessarily illegal sourcing. Stolen coins are another story.

It can also be said that since the Trump administration came into force, there is much less talk about mixers as illegal, and one gets the impression that mixers, as well as other coinjoin services, are under much less pressure than they were in the earlier period. It's as if crime has stopped.

[Lucius]

It can also be said that since the Trump administration came into force, there is much less talk about mixers as illegal, and one gets the impression that mixers, as well as other coinjoin services, are under much less pressure than they were in the earlier period. It's as if crime has stopped.

It seems to me that this is so only for the reason that "they" try to extract as much as possible from crypto while they are in power, so this idea of ​​mixing is not so problematic for them. You know how they say in our region "Steal, but also give to the people" (literally translated).

Therefore, all this so-called easing of measures may have its good side, but also its bad side, especially in the long term.

[d5000]

Has it ever been confirmed that that BTT account belonged to them? It didn't make any sense to me then (and it doesn't even make sense now) that such an agency would create an account on this forum to announce that it had seized a mixer.

I read that at least some have speculated that they were real indeed. I think it could make sense as we're here part of the "core" of the crypto community, and it doesn't cost them much to register and explain their stance on things.

However, allegedly from the beginning of next year, all members would have to be fully aligned with these laws and implement strict AML/KYC with all services related to cryptocurrencies.

I saw the list of types of services which require authorization according to MiCa some time ago. The original version of the mixer protocol I'm proposing is custodial, so it would for sure require KYC if it was launched in the EU or directed explicitly to EU customers.

However, a CoinJoin-based service like Samourai could be in my opinion launched in a way it's not custodial anymore and at the same time adhere to a similar protocol with pre-screening and limitation of the amount per day coming from the same origin. I remember having discussed this in another thread and my conclusion was that it was difficult but possible, having the list of services which have to be authorized according to MiCa in mind.

But anyway, one of the goals of such protocols would also to provide arguments for re-legalize non-KYC services, providing arguments that these don't help crime (to a greater extent than cash) while bringing potential benefits to society, with all the dangers associated to privacy violations and identity theft. It's currently not very likely that happens, but if crypto adoption starts to become stronger around the world I can see such an initiative happen. I  feel there is a slow resurging of this topic in the political discussion.

[OsaiEmma]

Mixers aren't exactly masking identity tied to a coin, it's only masking identity tied to an address. If we're more worried about the address than the coin, is that really censorship?

Censorship should be about id --> coin and not just id --> address, because it's during off/on-ramping your identity gets exposed and linked to a coin, and thus an address. So even after mixing, your id could still be linked to a coin.

All these headaches—for a legitimate user bothered about censorship—can be avoided by simply using a decentralized P2P service for off/on-ramping. Then who cares if the whole world sees where your coins have been to—unless you're doing something shady.

Now, if we know that mixers are there to mask your on-chain activities and not just about censorship, adding an AML check and restrictions probably defeats its purpose (i'd love to be corrected if I'm wrong).

The above is just my thought on the matter. But honestly, I love the filtering, it makes it less AML resistant and more AML compliant—which is one of the things holding mixers back, and labeled as facilitating laundry and coin theft.

[Synchronice]

Hasn't Wasabi wallet been bullied exactly for that? They pay a blockchain analysis company to analyze any UTXO which are trying to be coinjoined, which is a partnership and you offer the same idea. A Bitcoin mixer and chain analysis doesn't stand well together, really.

Also, I wonder, which jurisdiction will allow Bitcoin mixer to be run officially? Even in the form that you described? Offering people to improve their privacy and allowing scammers to analyze their coins before depositing to your mixer, so, you won't have to seize them in case something goes wrong, is legally a very terrible idea for any government. I mean, if it's really purely and honestly done this way, then it's really a terrible idea for any government.

[nc50lc]

In the protocol I'm proposing, the AML check is done even before the deposit, like an address check on a block explorer (e.g. Blockchair) but for the transaction/UTXOs. Or none at all is done, if the "small amount" rule is respected.

IICRC, Wasabi (specifically zkSNACK's coordinator) also did the same as mentioned by Synchronice above.
Still, they had been getting issues with regulations until they decided to shut it down.
But it may be different in today's standards and I see that you've already included KYC-related rule that the authorities will definitely like. (but most of us don't)

Anyways, it'll most likely receive the same negative criticism from privacy-oriented users.