{
  "description": "Chat with a model or serve it over HTTP in a container. Image, device, RAG, speculative-decoding and TLS options are left out.",
  "name": "ramalama",
  "subcommands": [
    {
      "name": "run",
      "description": "Chat with a model interactively.",
      "options": [
        {
          "names": [
            "--backend"
          ],
          "description": "GPU backend",
          "args": {
            "name": "backend",
            "isOptional": false,
            "suggestions": [
              {
                "name": "auto",
                "description": "Auto"
              },
              {
                "name": "vulkan",
                "description": "Vulkan"
              },
              {
                "name": "rocm",
                "description": "ROCm"
              },
              {
                "name": "cuda",
                "description": "CUDA"
              },
              {
                "name": "sycl",
                "description": "SYCL"
              },
              {
                "name": "openvino",
                "description": "OpenVINO"
              },
              {
                "name": "cann",
                "description": "CANN"
              },
              {
                "name": "musa",
                "description": "MUSA"
              }
            ]
          }
        },
        {
          "names": [
            "--name",
            "-n"
          ],
          "description": "Container name",
          "args": {
            "name": "name",
            "isOptional": false
          }
        },
        {
          "names": [
            "--ctx-size",
            "-c"
          ],
          "description": "Context size",
          "args": {
            "name": "ctx_size",
            "isOptional": false
          }
        },
        {
          "names": [
            "--temp"
          ],
          "description": "Temperature",
          "args": {
            "name": "temp",
            "isOptional": false
          }
        },
        {
          "names": [
            "--threads",
            "-t"
          ],
          "description": "CPU threads",
          "args": {
            "name": "threads",
            "isOptional": false
          }
        },
        {
          "names": [
            "--max-tokens"
          ],
          "description": "Max tokens",
          "args": {
            "name": "max_tokens",
            "isOptional": false
          }
        }
      ],
      "args": [
        {
          "name": "model",
          "isOptional": false
        }
      ]
    },
    {
      "name": "serve",
      "description": "Serve a model on a local port.",
      "options": [
        {
          "names": [
            "--backend"
          ],
          "description": "GPU backend",
          "args": {
            "name": "backend",
            "isOptional": false,
            "suggestions": [
              {
                "name": "auto",
                "description": "Auto"
              },
              {
                "name": "vulkan",
                "description": "Vulkan"
              },
              {
                "name": "rocm",
                "description": "ROCm"
              },
              {
                "name": "cuda",
                "description": "CUDA"
              },
              {
                "name": "sycl",
                "description": "SYCL"
              },
              {
                "name": "openvino",
                "description": "OpenVINO"
              },
              {
                "name": "cann",
                "description": "CANN"
              },
              {
                "name": "musa",
                "description": "MUSA"
              }
            ]
          }
        },
        {
          "names": [
            "--name",
            "-n"
          ],
          "description": "Container name",
          "args": {
            "name": "name",
            "isOptional": false
          }
        },
        {
          "names": [
            "--ctx-size",
            "-c"
          ],
          "description": "Context size",
          "args": {
            "name": "ctx_size",
            "isOptional": false
          }
        },
        {
          "names": [
            "--temp"
          ],
          "description": "Temperature",
          "args": {
            "name": "temp",
            "isOptional": false
          }
        },
        {
          "names": [
            "--threads",
            "-t"
          ],
          "description": "CPU threads",
          "args": {
            "name": "threads",
            "isOptional": false
          }
        },
        {
          "names": [
            "--port",
            "-p"
          ],
          "description": "Port",
          "args": {
            "name": "port",
            "isOptional": false
          }
        },
        {
          "names": [
            "--detach",
            "-d"
          ],
          "description": "Run in background"
        }
      ],
      "args": [
        {
          "name": "model",
          "isOptional": false
        }
      ]
    }
  ]
}