Msty

Msty began as the no-Docker local AI desktop and is now sold as an ecosystem of four products: Studio, a private AI workspace; Go, agents whose steps you review; Nexus, a local gateway over runtimes, providers, keys and usage; and Stack, versioned knowledge stacks. Studio carries the price list. It runs local engines - Ollama, Llama.cpp, MLX - next to online model hubs in one native app, with split chats, agent mode, knowledge stacks, and prompt, persona, skill and media studios. The tiers are Free at zero dollars for the desktop core, Aurum at 149 US dollars per user per year, which adds Studio Web, Azure and Bedrock providers, assistants and power-user features, a 349-dollar lifetime option, and a Teams plan from five seats with SSO, role-based access and audit logs.

The same page invites you to switch between Go, Nexus and Stack to compare what is available now, what is commercial, and what is coming next - so Studio is the only one with a number you can act on, and buying the rest means a sales call. Take it if local models should mean download-and-run: Free is genuinely bring-your-own, your keys and your Ollama, and the paid tier is a flat license rather than a second subscription stacked on your provider bill, with no token markup from Msty. Read the feature matrix first: much of what reads as the product is marked Limited on Free, and several headline capabilities sit behind Aurum. This is not the shape for a company-wide shared instance or a scripted API pipeline.