Topic: A Centralized Exchange With A Non-Custodial Wallet System - How feasible Is This (Read 300 times)

It also just works a lot better with Lightning (such as the already existing Robosats), since it has hodl invoices.
These allow users to 'lock in' a small collateral that ensures they stick to the protocol, without having to have 'virtual wallets' on the exchange or accounts to be managed and all of that.

If everything goes smoothly, the invoice is simply cancelled, if there's a misbehavior or scam attempt, it is accepted and the exchange keeps the collateral.

Keep in mind: Robosats basically is a centralized exchange with non-custodial wallet system. So one answer could be: 'It is feasible. It's already in practice'.

What you want is a "low-trust" platform if I understand well, where the exchange process itself is managed by the platform, but there are no "wallets" with funds the user needs to trust the platform.

It should be possible if the users sign transactions for their orders to the platform without broadcasting them.

If we have two users wanting to exchange 1 BTC for 23000 TetherUSD, the workflow would be that way:

1) User A creates a sell order for, e.g., 1 BTC. They sign a transaction of 1 BTC to an address of the platform, and sends the transaction to the platform without broadcasting it (can be made using the platform's website), and submit also an address to receive the Tether.
2) User B creates a buy order for the 1 BTC. They sign a transaction which sends 23000 Tether to an address of the platform, and sends the transaction to the platform without broadcasting it, and an address for the BTC.
3) Exchange checks both transactions for correctness and fees until allowing it to become orders on the platform.
4) If both orders match and the funds of both users are still unspent (see below for the problem of "abandoning" users), then the exchange platform broadcasts both transactions to the Bitcoin network, and use the funds to create transactions to the counterparty. User A's BTC minus fees are transacted to B's BTC address, and User B's Tether to User A's Tether (BTC/ETH/whatever) address.

Advantages/Disadvantages:

- With this setup only exact matches between orders are possible, otherwise the exchange would have to do some funds management themselves.
- Fees: Fees would tend to be higher than in traditional CEX's, because it would be more difficult for the platform to bundle transactions tooptimize transaction fees. This would only be feasible if the volume is high enough, otherwise there would be a high delay. On the other hand fees would be probably lower than in atomic-swap exchanges as there are less transactions involved and bundling may be sometimes possible.
- On the other hand, the setup would have a big advantage over Changelly-style exchanges as it would allow limit orders.
- The users can still try to double spend the funds to avoid an exchange. It would not be possible for them to run away with their counterparty's funds (like it is possible in Bisq), but they could prevent an unadvantageous exchange this way (although success would not be guaranteed). This is also a problem of atomic swap platforms.

I guess the niche for such a platform would exist but be rather small.

I think that it may be possible.
At the moment a "trade" is locked in, both buyer and seller have a fixed and short amount of time to send you (the exchange) the funds (they promised they exchange) and then you will send to each party its new funds.
Pretty much like an atomic swap, just it doesn't have to be "atomic".

The problem I see though is the if one part is not sending the promised funds in the amount of time, you'll have to send back to the other party what he sent, maybe even on your expense.
So it has to be tighter and I don't know how you can force somebody send funds from his wallet if he doesn't want to... unless you heavily interfere with his wallet and I guess that this is not wanted.

But with devs around you you may receive better ideas.

This is a beautiful idea of centralization and still retains customers privacy and control of finance. I have thought about this, though I'm not a tech guru, but what I could come up with is what NotATether said above.
Building an non custodial wallet and then incorporating an exchange in it. By this, everyone can exchange coin and save in same app. What you could find a bit difficult is the escrow system of the p2p.

Yes; this sounds to me a bit like Robosats. It's a centralized (onion) platform, but buyer and seller trade P2P and get their funds directly into their wallet. I may write a review fairly soon, in case anyone's interested. Anyhow; I noticed that the funds to sit in the service's wallet for a short amount of time, as it acts like an escrow. I don't believe they charge fees, but it would be possible to build something similar with a built-in fee, for sure.

For some inspiration, I'd encourage Fivestar4everMVP to just try it out once.
There's also this walkthrough that you can have a look at.

Another concept that I keep bringing up, which doesn't hand custody to the service, but instead uses multisig transactions is obviously Bisq.

My 2 cents about the opening statement: I found it funny how your main motivation seems to be 'building a very big business' and 'being a CEO to a big crypto currency exchange', instead of - I don't know - 'building a useful service' or 'creating a privacy-enhancing service', or maybe 'building something that helps Bitcoin'..
Sometimes, the best ideas come when you're not 'trying hard' to 'become a big CEO' or something like that; best example is satoshi. He wanted to create a trustless digital cash system, achieved it, and disappeared; probably gained little to nothing personally, but changed the world..
On the other hand, the modern-day altcoin creators are mostly scammers just out for 'big CEO' status and making lots of money quickly.

What do you think of ideas like localcryptos, as the money is in a guarantee that represents a multi-signature wallet and that wallet can contain more than 5 signatures so that the money can be spent without the intervention of the platform.
The difference between centralized and decentralized can be the database, if you have a large enough database and number of active users then it can be decentralized easily.

This means you want to build a wallet with a currency exchange system like Atomic, Guarda, OWNR, & Exodus.

If you ask me, that's a better idea. It requires virtually no funds aside from liquidity of various coins in the exchange system, you don't have to implement KYC (or evade it, thus running afoul of regulations) and you can take a commission from each currency exchange as a dev fee.

So, you should try to make a wallet with the currency exchange, with features contained only in wallets that don't have such an exchange feature e.g. MetaMask has DeFi browsing stuff, most BTC wallets have LN support, etc.

If centralized only means there's a place (usually website which controlled by a group/company) to meet buyer and seller, it's possible. But people usually call it P2P rather than CEX even if both buyer and seller need to use that website. Otherwise, i don't see how it's possible on technical level.

If the concept could work it will be a nice one but unfortunately, the mode of operation of Centralization exchange doesn't support the use of non-custodial wallets, and the best bet is to offer a non-custodial wallet system on crypto exchange is still through an automated market maker type of decentralized exchange.

Hi Guys.

So for a very long time since I got into crypto, I've always wanted to build a very big business that's powered by crypto currency, and to this effect, I've nursed the idea of being a CEO to a big crypto currency exchange, this idea have been with me since 2017 and in 2020, I brought together some group of friends consisting of mostly developers, programmers and graphic designers, they actually bought the idea after i discussed it with them, and then, we agreed to work together to actualize the dream.

But unfortunately, we gave up even before we could put together something meaningful, what was the reason?

1. There were already a lot of exchanges and from our research, over 98 percent of this exchanges are not doing well, they extort money from their users as a means of making good money required to keep the exchange running, and this is something we don't wanna do.

2. We couldn't find anything unique in our proposed exchange whitepaper, every feature we planned to implement were already available on almost all the exchanges that are currently in existence, and like I said before, 98 percent of this exchanges weren't and aren't doing well, So we assumed that ours will perform even more badly because we plan to start it in a very small way due to very little funds available for our usage.

So after deliberating on the above and many other issues, we decided to give up, majorly because we lacked ideas, we needed some feature which will make our exchange unique and at that time, we had none.

So just a day before yesterday as I was browsing through this forum, the idea on the subject of this thread came, so I decided to find out here, how possible/feasible it is to build/run a centralized exchange with full non-custodial wallet system implemented?

In my own personal observation, I don't think its possible to run a centralized exchange and don't have access to users wallet, but then, somebody here might know something I don't know, so all comments, observations, suggestions, opinions are highly welcome.