AI 101

llama.cpp server

Also known as: llama server · llama-server

Local

llama.cpp server is the HTTP server bundled with llama.cpp. Start it and your model is exposed at a local endpoint that mimics the OpenAI API.

  • GET/POST /v1/chat/completions
  • Drop-in for tools that expect an OpenAI-compatible API.
  • Zero cloud, zero account.

Why it matters: point any OpenAI-style client at http://127.0.0.1:8080 and it just works. This is how you get a "cloud-like" API on your own machine.