[{"data":1,"prerenderedAt":114},["ShallowReactive",2],{"run-qwen-locally":3},{"ollama":4,"llama-cpp":52,"vllm":74},{"id":5,"slug":5,"name":6,"url":7,"curatedDescription":8,"category":9,"section":14,"repo":18,"media":43,"source":46,"sync":50},"ollama","Ollama","https://github.com/ollama/ollama","Dead-simple local LLM runner with a one-line install, model registry, and OpenAI-compatible API.",{"id":10,"slug":10,"title":11,"navTitle":12,"order":13},"inference-engines-and-serving","Inference Engines & Serving","Serving",3,{"id":15,"slug":15,"title":16,"order":17},"local-on-device-inference","Local / On-device Inference",1,{"host":19,"owner":5,"name":5,"fullName":20,"stars":21,"license":22,"primaryLanguage":23,"homepage":24,"topics":25,"description":40,"pushedAt":41,"archived":42,"fork":42},"github","ollama/ollama",173169,"MIT","Go","https://ollama.com",[26,27,28,29,30,31,32,33,34,35,36,37,38,5,39],"deepseek","gemma","gemma3","glm","go","golang","gpt-oss","llama","llama3","llm","llms","minimax","mistral","qwen","Get up and running with Kimi-K2.6, GLM-5.1, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.","2026-06-04T18:05:37Z",false,{"iconUrl":44,"avatarUrl":45},null,"https://avatars.githubusercontent.com/u/151674099?v=4",{"readmeUrl":47,"readmeLine":48,"sourceHash":49},"https://github.com/alvinreal/awesome-opensource-ai/blob/main/README.md",269,"fb157d63a81e09435737672cb2c10cb3ef440c1b9be35512bd9443bd8cb68959",{"lastSyncedAt":51},"2026-07-12T12:00:53.360Z",{"id":53,"slug":53,"name":54,"url":55,"curatedDescription":56,"category":57,"section":58,"repo":59,"media":69,"source":71,"sync":73},"llama-cpp","llama.cpp","https://github.com/ggml-org/llama.cpp","Pure C/C++ inference engine with GGUF format support. The gold standard for CPU/GPU/Apple Silicon on-device running. Includes llama-server for OpenAI-compatible API. Now at 100K+ stars.",{"id":10,"slug":10,"title":11,"navTitle":12,"order":13},{"id":15,"slug":15,"title":16,"order":17},{"host":19,"owner":60,"name":54,"fullName":61,"stars":62,"license":22,"primaryLanguage":63,"homepage":64,"topics":65,"description":67,"pushedAt":68,"archived":42,"fork":42},"ggml-org","ggml-org/llama.cpp",114610,"C++","https://llama.app",[66],"ggml","LLM inference in C/C++","2026-06-04T17:13:26Z",{"iconUrl":44,"avatarUrl":70},"https://avatars.githubusercontent.com/u/134263123?v=4",{"readmeUrl":47,"readmeLine":72,"sourceHash":49},268,{"lastSyncedAt":51},{"id":75,"slug":75,"name":76,"url":77,"curatedDescription":78,"category":79,"section":80,"repo":84,"media":109,"source":111,"sync":113},"vllm","vLLM","https://github.com/vllm-project/vllm","State-of-the-art serving engine with PagedAttention and continuous batching. Currently the fastest production-grade LLM server.",{"id":10,"slug":10,"title":11,"navTitle":12,"order":13},{"id":81,"slug":81,"title":82,"order":83},"high-performance-serving-and-api-servers","High-performance Serving & API Servers",2,{"host":19,"owner":85,"name":75,"fullName":86,"stars":87,"license":88,"primaryLanguage":89,"homepage":90,"topics":91,"description":107,"pushedAt":108,"archived":42,"fork":42},"vllm-project","vllm-project/vllm",81935,"Apache-2.0","Python","https://vllm.ai",[92,93,94,26,95,96,32,97,98,33,35,99,100,101,102,103,39,104,105,106],"amd","blackwell","cuda","deepseek-v3","gpt","inference","kimi","llm-serving","model-serving","moe","openai","pytorch","qwen3","tpu","transformer","A high-throughput and memory-efficient inference and serving engine for LLMs","2026-06-04T18:01:24Z",{"iconUrl":44,"avatarUrl":110},"https://avatars.githubusercontent.com/u/136984999?v=4",{"readmeUrl":47,"readmeLine":112,"sourceHash":49},286,{"lastSyncedAt":51},1783865043227]