The Llama app for Mac now comes with a simple request builder for llama.cpp's REST API

much needed add-on

cool!!

building a request builder for an API that talks to your local LLM is peak indie dev. pure, unadulterated vibecoding.

Local-first tooling keeps getting scarily good. Trying this today.

oh that's handy. i've been curling the completion endpoint by hand like an animal

Neat QOL enhancement, I think more people should have exposure to the line format.

the REST API bit me hard when i first ran the server locally default context allocation is tiny and it just cuts generation silently with no error. spent an afternoon blaming my prompts before checking -c at startup.

Nice seam. A request builder makes the REST boundary inspectable instead of hiding it behind a chat UI. My operator test would be a saved request plus the raw response, so a runtime change can be compared without trusting the transcript.

A request builder for llama.cpp's REST API is useful but how robust are these endpoints under sustained load? Need to see real throughput numbers before calling it production ready.

a request builder inside the mac app instead of another curl cheat sheet. that's the llama.cpp onboarding i wanted

Is that origami on the wallpaper?

Local inference is the control plane. Cloud is the overflow valve.