RKLLM provides tools to convert and quantize trained AI models into RKLLM format on a PC, then deploy and run them on Rockchip development boards using a C/C++ runtime API. It supports a wide range of LLMs including Qwen, Llama, ChatGLM, Phi, Gemma, InternLM, and vision-language models, targeting Rockchip SoCs like RK3588, RK3576, RK3562, and RV1126.