[ANN]: Bopwire: The Tokenized P2P Music Sharing App - Attention/PoW chain

[nyantheory]

This is a proof of concept. The PoW and validation loop are built around song fingerprints, unique songs build a block there are both elements of PoW and proof of attention for both block generation and minting.

lots of work still to do on it, but it sort of works. I will update the thread as features are fleshed out.

I am not looking for investors at this time, but if anyone needs a free local music player for their phones please try my software.


Windows and Android binaries
https://github.com/knobcore/musicchain/releases/tag/v0.1.0-20260610

source code
https://github.com/knobcore/musicchain

[nyantheory]

core technology and backend mostly working now

the project has been changed to bopwire. if a mod can change the title that would be great

i managed to get a premium domain for this project
https://bopwire.com/

here is a web player note: the songs are actually stored on my PC's player/fingerprints were added to the blockchain and being streamed through the mini node via DHT of the data matching content hash for seeding. i'm hosting a streaming service that connects to a p2p network on a github page it's pretty magical
https://bopwire.com/player/

build scripts for most systems to run nodes etc is in my github
https://github.com/knobcore/bopwire

whitepaper (i'm going to bash together a real one over the next two days off this is just the "hey spit out the technical implementation" version)
https://bopwire.com/whitepaper.html

[NotATether]

I mean, you could definitely do better with the website.

You're free to message me if you're interested in that.

[nyantheory]

I mean, you could definitely do better with the website.

You're free to message me if you're interested in that.

Of course it is. I don't deny that. To be fair, I haven't even touched front end, or auto playlist generation, or anything like that. I've been spending the last several weeks getting the network into a usable state, and still a lot of the things in the backend is not tested yet. The fact that I can pull audio to the website at all from a single seeder behind restrictive NAT is a milestone in that respect.

There is no block explorer, there's no security for seed phrases they're just plain text on disk for now, little DSP for audio playback (which I'm going to put an EQ and a limiter in at least - this is my area of expertise, not so much cryptography) songs.list returns all the songs not just a subsection of what you're exploring (which needs to be optimized). All of these things have to be done before i even touch the front end more.

However, if someone wants to help develop it they are free to do so. I am looking for developers, but you have to be better than what I can generate with Claude. Like "I built a SHA-256 miner and wallet in x86 assembly in 2015" level shit before I accept your commits.

It is an open source project. I don't want to start a company around it, but am open to a 501c3, or maybe i'll just let it run wild without all of that.

The only thing I might make closed source is DMCA management, though currently it isn't though your wallet has to have a mod flag to login. I am also worried about escrow security, which is a mod action and is why that may be the case. If someone knows of a system that does this please point me to examples.

[NotATether]

However, if someone wants to help develop it they are free to do so. I am looking for developers, but you have to be better than what I can generate with Claude. Like "I built a SHA-256 miner and wallet in x86 assembly in 2015" level shit before I accept your commits.

Hell yeah!  You're talking to the right person for this.

My telegram is @Zenul_Abidin for detailed discussion.

I'll get a discount just for you

[nyantheory]

However, if someone wants to help develop it they are free to do so. I am looking for developers, but you have to be better than what I can generate with Claude. Like "I built a SHA-256 miner and wallet in x86 assembly in 2015" level shit before I accept your commits.

Hell yeah!   You're talking to the right person for this.

My telegram is @Zenul_Abidin for detailed discussion.

I'll get a discount just for you

To be clear, the last thing that's going to be worked on is the website. All of the things I mentioned have to be implemented before I even am going to worry about it. That's at least 2-4 months of work, even with AI assistance for a solo developer to implement all of what I need before the website is updated. Not to mention a good website might attract an investor but i've seen a lot of good websites and lack of tech and a sort of "throw shit at the wall to see what sticks" mentality generally in the crypto space with dead githubs and the rest which I want to avoid. And considering that I'd have to choose between my pro max plan and paying for web dev I'm obviously going to choose that, especially with Fable being dropped again today.

And I need someone to explain to me why using a fingerprint is not complex enough to use as a point of failure in validation checks against history, and any attack vectors that may show up from that. Blocks don't mint in this project, only validated plays that are signed by mini nodes that verified that it relayed the data, and it passes all sanity checks and validated by full nodes as real can mint - ie if more than one play comes from the same MAC, device ids, and OS it's flagged by the full node as invalid. If the mini node doesn't relay the bytes it won't sign the validation either to mint and reject the play from the count. In that sense, the bot would have to decode the audio in real-time. A threadripper could probably do a thousand decodes or more at once, but it won't allow it from the same ip and hardware fingerprint you'd also have the overhead of many VMs which go through your ip anyway and good luck changing the vendor and device ids. I'm concerned about that still as well. It's a novel idea to use unique fingerprints as a clock to move a chain forward, and honestly it's such a good idea if nobody can explain why that is that it will probably be copied ad infinitum, because you could do the same thing for movies or anything else that can be fingerprinted with FFT analysis that is a consumable. The computation is just lean enough to be a green PoW structure, but the cost structure is still there at scale.

[NotATether]

Understood. If you ever need one in the future, you know who to call.   

[nyantheory]

Understood. If you ever need one in the future, you know who to call.  

Of course, your work looks good.

What I really need right now is for someone with a large mp3 library to run the player (not web version) and add it so the network sees it. That's really it at the moment. Fingerprinting is dead slow though. My phone takes about 30 minutes to fingerprint this small library. I wonder if I could have Claude port chromaprint to opencl to speed that up a bit.

[nyantheory]

did some work on auto playlist generation in discover

added album art scraper

looks more like a Spotify clone now

https://bopwire.com/player/

which also carries latest app releases

[nyantheory]

Added slak and napstr integration you can now download songs from these networks and when they're added to your library they automatically appear on chain

when you search a song from these networks you can also preview them without committing to library. there is a 1 hour cache for previewed songs

https://x.com/laincobayne/status/2092864733229822415

for a video of progress

[nyantheory]

Block explorer!

https://bopwire.com/player/explorer.html#/blocks