| Item | URL | WHY |
|---|---|---|
| llama.cpp | ggerganov/llama.cpp | Running local models and building tools around it. |
| various open source models, most recently mistral-medium 3.5 and mistral-small 4 | medium, small | Currently enjoying running these models locally in a Build/Plan configuration - both open source! |
| building an openai-compatible API which injects its own tools | rubyllm | being able to inject a harness with new tools seems very useful, early development stage. using rubyllm to develop |
| Seabird | getseabird/seabird | native gui k9s? yes please. |
| Knative | https://knative.dev | It's kind of cool to look at k8s as "just" the platform rather than something like heroku/render/fly.io |
| Ministack | ministackorg/ministack | local AWS emulation without the localstack account! |