# Chaty

A private desktop app for chatting with, and building with, open models that run entirely on the user's own machine, combining chat, a coding agent, image and music generation, and document Q&A.

Chaty runs GGUF models through llama.cpp and, on Apple Silicon, MLX models natively, alongside image generation through stable-diffusion.cpp and music generation through audio.cpp, all without an account, cloud connection or telemetry. Its coding agent is built specifically for small models that fit on a laptop, checking edits against the file while they are still being written and formatting tool calls the way each model was trained to expect, so a small local model can complete real coding tasks.

Chat responses stream in with rendered code, tables, math and diagrams, reasoning shown in a collapsible panel, and a live-preview canvas for generated pages. A model store searches Hugging Face and reports which files fit in the user's available memory, and all models, chats, documents, pictures and music are kept in a folder on disk that can simply be deleted to remove everything.

Website: https://chaty.ca