defilantech/LLMKube
Kubernetes operator for self-hosted LLM inference across a heterogeneous GPU fleet: NVIDIA CUDA, AMD Vulkan, and Apple Silicon Metal. Runtimes: llama.cpp, vLLM, TGI, mlx-server. Multi-GPU sharding, model caching, OpenAI-compatible endpoints. Apache-2.0, run across homelab and on-prem fleets, actively developed.
Star count as of 02:27 (BKK).
defilantech/LLMKube has 245 stars and gained +27 of them in the last 30 days, a 12% change. It is written in Go, grouped under ai-infra, released under Apache-2.0. A merged pull request waits a median of 6h from opening to merge, measured on 189 pull requests. Last pushed on 27 Oct 2026.
Star history
Loading star history…
Growth
| Period | Stars gained | Growth | From | Measured |
|---|---|---|---|---|
| Last 6 hours | — | — | — | Not measured |
| Last 12 hours | — | — | — | Not measured |
| Today | 0 | 0.0% | 245 on 26 Oct 2026 | Daily snapshots |
| Last 7 days | — | — | — | Last measured 25 Oct 2026 |
| Last 30 days | +27 | 12.4% | 218 on 27 Sep 2026 | Weekly history |
| Last year | — | — | — | Not measured |
A period with no current measurement is left blank: a delta needs a measurement at both ends of the period. The 6- and 12-hour periods need two intraday measurements exactly that far apart, the day-and-longer periods need snapshots on both dates, and the last year also needs star history from a year ago, which we keep for one year.
Embed this badge
Copy either snippet into the repository's README. The badge shows total stars and the gain over the last 7d (the same window as the one below) and is hotlinked directly, with no account or token.
[](https://bestofgit-blue.vercel.app/projects/defilantech/LLMKube)
<a href="https://bestofgit-blue.vercel.app/projects/defilantech/LLMKube">
<img src="https://bestofgit-blue.vercel.app/badge/defilantech/LLMKube.svg?window=7d"
alt="Star and growth badge for defilantech/LLMKube on BestOfGit" />
</a>
