Bring Your Own Local AI Model (v3.3)

Run any GGUF language model locally inside Hedy. Gemma, Qwen, Mistral, and other models from Hugging Face all work. Drop a model file in via the file picker and Hedy uses your hardware, fully offline, 100% private.

Model metadata (parameter count, quantization) shows on each custom model card so you can pick the right one for your device.