{
  "schemaVersion": "1.0",
  "updates": [],
  "companyNews": [],
  "id": "b636",
  "language": "en",
  "name": "llama.cpp",
  "originalName": "llama.cpp",
  "library": "tools",
  "category": "Model development",
  "categoryKey": "模型开发",
  "categoryKeys": [
    "模型开发"
  ],
  "categories": [
    "Model development"
  ],
  "entryPoints": [],
  "summary": "Run local models in C/C++, with command-line tools, quantization utilities and a model server with a web interface.",
  "introduction": "llama.cpp is a model-inference project maintained by ggml-org, not a single Llama model. With compatible model files, users run a chat CLI or start a server with a web interface and compatible APIs. The project supports GGUF workflows, quantization and multiple CPU/GPU backends for local or self-managed inference. Check model files, licenses and hardware support separately; the project name does not imply every model is compatible.",
  "capabilities": [],
  "company": {
    "name": "ggml-org",
    "originalName": "ggml-org"
  },
  "officialUrl": "https://github.com/ggml-org/llama.cpp",
  "profileUpdatedOn": "2026-10-22",
  "sources": [
    {
      "kind": "product",
      "url": "https://github.com/ggml-org/llama.cpp"
    },
    {
      "kind": "documentation",
      "url": "https://github.com/ggml-org/llama.cpp/tree/master/tools/server"
    }
  ],
  "news": [],
  "links": {
    "interactive": "products/llama-cpp--b636.html",
    "reading": "read/en/products/llama-cpp--b636.html",
    "markdown": "read/en/products/llama-cpp--b636.md",
    "json": "read/en/products/llama-cpp--b636.json"
  }
}