Share your local, premium and WebGPU models with anyone with just a link! created by @jasonkneen
$MOLTRATS
5FXbnqTjetSzdn9uVLPCQXCGf8iYo5WW8bsrQJCVBAGS
Moltrats are alive and well and networking -- we're going open-source which means we'll be powering many more projects that need peer to peer inference!
The agentic OS only needs basic primitives -- in this case we have the chat, files, and a renderer and with that I've created a concept character, then generated a sprite sheet, then asked for an app to cut up and manage the frames.
All running locally.
I'll bin that app after this because the prompt is saved and it can recreate it better next time.
This is STILL broken of course -- i should not have to type or move things around -- voice should be all it needs.
Thats coming
Top requests we're already working on:
- private model sharing (we do already with resources but this is coming)
- dashboard reporting -- we have a snazzy dashboard coming that's fully customisable and reward system
So getting a lot of interest on @moltrats and it looks like coins have been setup on various markets.
So far I have looked into the @bankrbot community and for this project I think it's a really good fit and means I can develop this into something bigger with the fees.
This doesn't mean I'm abandoning @huggiapp as @moltrats is all about open-source free models and Huggi is about using those models so they solve two different problems and fit perfectly together.
This is the @moltrats chat -- you can use the autorouter and just fire way it's all free. Or you can pick specific rats that are online. Rats can also share tools, skills and context as part of their model so there's LOTS of expansion.
Currently working on the rewards system which is gas for gas so the more you share the more gas you earn and the more you can use so if I share 2 hours of my Opus usage, and you use it, I earn gas and can use that when I need it.
Working through it all but it's going to be epic.
Ok so what is https://t.co/Y5YDcvx7yy all about?
It solves a problem that's a common one. Sharing inference fast. I wanted to share my DGX sparks with someone; normally that means opening up ports; setting up tunnels etc. PITA
I wanted a better way...
1/3
Introducing Moltrats the first peer to peer inference sharing platform for humans, agents, and digital crustaceans.
Share webGPU, local or cloud models over the web securely via WebRTC, MCP and/or REST APIs to any agent or human with nothing to install. Just open a browser or URL endpoint.
๐งต 1/6