NIBIA Fabric
github.com/nibia-ai/fabric- Category
- AI Tools
- Rank
- No. 2007Tools index
Previous survey · No. 1908 ·
- Pricing
- Open Source
- Type
- TOOL
- Use case
- Models: Train & Run
- GitHub
- 6 stars
- Latest release
- v0.7.0-alpha
- Date
About
NIBIA Fabric is a distributed LLM inference system that coordinates computers on the same trusted LAN and partitions GGUF model execution across them. Larger models can use the combined CPU and RAM of several machines instead of one. It builds on llama.cpp and runs on macOS, Linux and Windows. The current release is an experimental alpha.
At a glance
NIBIA Fabric splits local language-model inference across computers on the same network. The experimental release is worth exploring if a model needs more memory than one machine has and you can use spare CPU and RAM on nearby machines.
Good fit
Local inference experimenters with multiple supported machines on a trusted LAN.
Consider something else when
Skip it if you need production stability, a public network deployment or a guaranteed speed improvement.
Start here
Use two test machines on a trusted wired network. Run a small GGUF model on the primary, then inspect the execution plan for a larger model and check whether it assigns work to the second machine.
Cost and requirements
Apache-2.0. The alpha targets Apple Silicon macOS and x86-64 Linux and Windows. You supply the machines, model files and network. The documented macOS alpha archive is unsigned.
Limits to know
The primary serves its API on loopback by default. Adding machines can make a larger model fit, but network overhead affects load time and generation speed. The maintainers report different results on Wi-Fi and Ethernet for the same setup.
Sources checked 2026-11-05. Source-based assessment.
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.