Managing Models

Model Hub is the central place to connect providers, install local models, compare model options, and tune model behavior.

Model Hub Workflow

For most users, model setup follows this order:

1. Open Model Hub from the sidebar or Command Palette
2. Connect an online provider, install a local model, or do both
3. Monitor active service and model downloads from the Model Hub badge
4. Set model defaults and purpose tags
5. Tune model parameters or apply a preset when needed

Model Hub Sections

• Model Providers: Connect and manage hosted providers and their models
• Local Models: Discover, install, import, and manage models across Ollama, MLX, and Llama.cpp
• Model Squad: Assign preferred models to specific tasks
• Model Matchmaker: Get guided recommendations based on your needs
• Cost Calculator: Estimate hosted model usage costs
• VRAM Calculator: Estimate whether local models fit your hardware
• Context Explorer: Learn how context windows and token usage affect conversations

Active Downloads

When a model or local service is installing, the Model Hub sidebar icon displays an activity badge.

• Select the badge to return to Model Hub
• View active local and remote install counts
• Open the active installs panel to review progress or cancel an installation

Online and Local Models

Msty Studio supports two model sources:

• Online Providers: Hosted APIs that you connect with API keys
• Local Models: Models you install and run on your own machine

Online Providers

Use online providers when you want quick access to hosted models and managed infrastructure.

Go to Online Providers for supported providers, setup steps, and OpenAI-compatible bring-your-own endpoints.

Local Models

Use local models when you want more control over privacy, offline access, and local performance tuning.

Go to Local Models for the unified Featured, Installed, Model Search, and Import workflows.

Find the Right Model Faster

Use Model Matchmaker when you are unsure which model to start with.

Organize Models for Daily Use

Set a Default Model

In the conversation model selector, click the star icon to set a default model.

Rename Models and Providers

In Model Hub or the conversation model selector, click the edit icon to rename a provider or model.

Set Model Purpose Tags

Purpose tags help categorize what each model is best at.

• Text: General text generation and understanding
• Coding: Code generation, analysis, and debugging
• Tools: Tool/API usage with Toolbox (MCP)
• Image: Image generation models (see Image Generation)
• Vision: Image understanding and analysis
• Embedding: Retrieval and vector workflows, including Knowledge Stacks
• Streaming: Near real-time data processing
• Thinking: Reasoning models with extended thinking controls

Provider and Model Logos

You can change provider and model logos from the edit dialog in Model Hub.

Related Resources

Videos

Blog Posts

Unleashing Local AI - Transform Your Coding Experience with Msty

Discover how to enhance your coding workflow using local AI models with Msty, GitHub Copilot, and Roo Code for personalized, offline support.

The AI Settings Dictionary

AI settings are a powerful way to customize your AI experience. Here is what they all mean and how to use them.

5 Ways to Use Local Inference with Msty Studio

Explore five practical ways enterprises can run local AI inference with Msty Studio, keeping data private while giving teams a powerful, easy-to-manage front end for productivity.

5 Reasons You Should Try MLX with Msty Studio

Discover how MLX and Msty Studio unlock fast, local AI on Apple Silicon Macs, letting you create, experiment, and explore powerful models effortlessly.

Powerful, Local Inference with LLaMA.cpp

A focused guide to using LLaMA.cpp in Msty Studio, including quick setup, key settings, and why LLaMA.cpp is a strong local inference engine.