HN Explorer

Show HN: Nibia Fabric – Turn everyday computers into a shared AI compute cluster (github.com)

1 points by abelop ↗ HN
I built NIBIA Fabric, an open-source distributed LLM inference system that pools CPU and RAM across macOS, Linux, and Windows machines.

It builds on llama.cpp, with capacity-aware scheduling, adaptive memory reservation, persistent tensor caching, and an OpenAI-compatible API.

I’ve validated the current alpha on three physical machines across several models, including a 30B-class Qwen model, as well as GPT-OSS 20B. Feedback on the architecture and use cases is very welcome.

4 comments

[ 64.8 ms ] story [ 622 ms ] thread
kickstartaioffi
very interesting
abelop
Thanks! Happy to answer any questions about how it works or the architecture.
H4kken
Sounds pretty cool, do you have the link to the github so we could take a look ?

Also, who do you plan on targeting with this promising project? like idk if the average ai user nowadays would like to bother with self hosting everything, but i might be wrong

i feel like a more enterprise oriented scope could interest more ppl