Best for this computer
The models page opens with a pick per activity - coding with you, everyday chat, seeing images, health questions - each chosen from what your machine can actually run. No guessing, and no downloads that were never going to fit.
Model management without a terminal
Browse, download, and switch open models in the app, with hardware-fit guidance - what runs fully on your GPU, what will be slower, what won't fit. Models that can drive project work carry an Agentic label, so you know before you download.
Every model says who made it
Model cards name the maker of the weights, who packaged the files you download, and for community builds, what they are based on and how they differ. You always know whose work you are running.
Engines for your hardware
Vulkan and Metal out of the box, and a one-click CUDA engine for NVIDIA machines - the app picks safe defaults and steps down gracefully if your hardware protests. More than one GPU? Bigger models spread across them.
Bigger than your graphics card
Mixture-of-experts models can run with their always-on layers on the graphics card and their experts in system memory - so a model well beyond your card's size still runs at a good pace. The app knows which models split well and grades them honestly.
Models live where you say
Choose where model files are stored - a second drive, an external disk - and downloads follow, with free space checked before every download and existing models moved safely.
Apple Silicon MLX engine (preview)
Macs with Apple Silicon can add an optional MLX engine, then fetch MLX versions of supported models - chats run on MLX while everything else stays on the standard engine. Whether it is faster depends on your Mac; nothing changes unless you install it.
Online frontier models (paid plans)
Frontier models from every major lab - an optional paid service you sign in to with your Flowsta Identity. A monthly allowance, fair metered pricing, every price shown up front. An add-on, never a dependency.
