Online Providers
Connect hosted AI providers and manage cloud model access
Online providers let you use hosted models with your API keys, without running model inference on your device.
Need API key links? See Find Online Model API Keys.
Connect an Online Provider
Open Model Providers
Go to Model Hub > Model Providers.
Add Provider
Select Add Provider and choose your provider.
Enter Credentials
Add your API key (and any required endpoint details), then save. If you need key locations, use Find Online Model API Keys.
Test in Chat
Open a conversation and select one of the provider's models.
Supported Providers
Msty Studio supports these online provider options:
- Anthropic
- Cohere
- Gemini
- Groq
- Mistral AI
- OpenAI
- OpenRouter
- Perplexity
- Together AI
- SambaNova
- xAI
- Azure OpenAI
- Amazon Bedrock
For key locations and endpoint references, see Find Online Model API Keys.
Bring Your Own Provider
If your provider exposes an OpenAI-compatible API, you can connect it by using an OpenAI-compatible endpoint in Model Providers.
This is useful for:
- Self-hosted gateways
- Enterprise AI platforms
- Private inference endpoints that follow the OpenAI API format
Manage Provider Setup
After connecting, you can:
- Rename providers and models for readability
- Set a default model in chat
- Apply model purpose tags from Managing Models
- Apply model-parameter presets, including supported provider-native tools
Provider-Native Tools
Some hosted providers expose tools directly through their model APIs. These tools run through the provider rather than through an Msty Toolbox toolset.
Msty Studio includes model-parameter presets for commonly supported native tools:
| Provider | Included native-tool presets |
|---|---|
| OpenAI | Web Search, Code Interpreter |
| Anthropic / Claude | Web Search, Web Fetch, Code Execution |
| Google Gemini | URL Context, Google Search with URL Context |
| xAI | Responses API with Web Search |
| Groq | Browser Search |
Apply a Native-Tool Preset
Select a compatible model
Open a conversation and select a model from the supported provider.
Open Model Parameters
Select the model-parameters button beside the model selector.
Choose a preset
Open Presets and select the provider-native tool preset, such as OpenAI with Web Search or Claude with Code Execution.
Apply the parameters
Select Apply, then send a prompt that can use the enabled capability.
When a response uses a native tool, supported tool results and grounding sources appear with the response.
Native-tool availability depends on the provider, model, API account, region, and current provider capabilities. A preset can be available even when a specific model does not support every configured tool.