Hello. Bitcoin is an untrusted medium. And with the advent of Lightning, many began to move away from this principle towards the use of custodians (centralized solutions). I thought about it after the LN bot was closed and there were problems with withdrawals from users.

What I suggest

There is an option with an automatic software storage, where bitcoins are stored without movement, which allows you not to spend commissions. It can be a specially generated address to which no one has access except the users of this repository.
But at the same time, the user works with him with his bitcoin key. And the node is only a client (interface) that provides a convenient format for interaction: speed + transfer by login.
In fact, the scheme is as follows:
  1. Users send transfers to others.
  2. They or the service itself collect a bundle of user transactions, and push them into the Bitcoin network with one transaction.
  3. The repository moves funds, if necessary, or simply fixes the state that BTC or other tokens are with another user.
Roughly speaking, a Client can collect 1000 user transactions; throw them into storage, spending $ 7 (an example is simple - the transaction is more expensive because there is data), and users pay only 0.7 cents for the transaction.
At the same time, if the Client closes, the user can transfer information to another service through his wallet / standardized backup, and he will start transactions in the storage for a meager extra. payment. Or the user can throw his own transactions, of which there are not a thousand since the service scam, but two or three, spending $ 1-2.
Something like that.

Disadvantages:

As they say, every solution can be less secure/productive/decentralized or cheap. In my version, it may well turn out that at times of large commissions in the Bitcoin network, the commission for adding a state to the repository may exceed $ 7, which is why the final commission from users of the service may not be 0.7 cents, but, for example, 10 cents. The same problem may arise with the usual state of commission prices, if I did not foresee in my idea that the cost of sending a transaction to the repository with a state package for 1000 transactions would cost > $ 7, for example, $ 20, 50 or 100.
In this case, the final cost for users will have to rise in price, but this is a small fee for maintaining decentralization and lack of trust...

Signature issue

As I wrote at the beginning, the user signs transactions by interacting with the repository with his Bitcoin key. This means that it will not be possible to use the interfaces of services that help to interact with the repository: only popular Bitcoin wallets. But another option would be to create a special key to which the service will have access, but which will not give access to the main funds in the Bitcoin wallet. At the same time, such a key will be generated locally based on the user's private key. It turns out a kind of second-level key for an untrusted environment, and then it will be possible to use third-party wallets from services, and not just Bitcoin wallets.

Terminology

I hope I didn't confuse you with services... Simply put, these are LN nodes, like getalby.com , Lightning bots, etc.

Everything

Thank you for your attention. I will be glad to comments and likes. I wonder what you think about my idea? And maybe something like this is already in development / done? Thank you.
Have a look at statechains idea by Ruben Somsen. There is an implementation by Mercury wallet as well.
You practically describe UTXO pool controlled by the service. However it is non custodial because users have backup presigned transactions so they could redeem coins onchain.
reply
Yes. I've read about them. Then somehow it should be combined with Lightning, and it will be great!
reply
Welcome here
reply