kdeps-recipes-llama-server

id: llama-server
name: llama-server (GGUF)
description: llama.cpp llama-server serving GGUF weights over OpenAI-compatible /v1
version: 1.0.0

api:
  port: 8000
  base_path: /v1
  chat_path: /v1/chat/completions
  health:
    method: GET
    path: /v1/models
  auth:
    mode: none
    env: LLM_API_KEY

engine:
  kind: llama-server
  base_image: "ubuntu:24.04"
  packages: []
  install: |
    # llama-server binary is resolved/copied by the image builder at build time
    true
  command: ["llama-server", "--host", "0.0.0.0", "--port", "8000"]
  env: {}
  internal_port: 8000
  openai_bridge: false

models:
  strategy: copy
  default: ""

resources:
  gpu: optional
  memory_hint: "8Gi"

capabilities:
  - chat