In `modelConfigurations` introduce a new property for configuration name

[waldekmastykarz]

Prerequisites

• I have checked the latest documentationand this feature doesn't exist
• I have searched for similar feature requests and found none
• I have linked to related issues or discussions in the description (if any exist)

Type of Improvement

New Functionality

Proposed Solution

Extend modelConfigurations with a new property for the configuration name:

{
    "prompty.modelConfigurations": [
      {
        "name": "Phi-4 (Foundry Local)",
        "model": "Phi-4-mini-instruct-generic-gpu",
        "type": "openai",
        "api_key": "local",
        "organization": "",
        "base_url": "http://localhost:5272/v1"
      }
    ]
}

Today, when you specify model configurations in the prompty.modelConfiguration, you have to use the name property to specify the model's name. Since configurations are matched by name, it makes it impossible to have different configurations for using the same model on different hosts (ie. Ollama vs. Foundry Local vs. Azure).

[leestott]

@sethjuarez this is a feature we would like for https://github.com/microsoft/foundry-Local/ @samuel100

[sethjuarez]

Good call - I will see about adding invokers for this.

[sethjuarez]

Resolved in the v2 VS Code extension. The v1 extension used modelConfigurations in settings.json where the name property was overloaded as both the display name and the model identifier, making it impossible to have multiple configurations for the same model on different hosts.

The v2 extension replaces modelConfigurations entirely with a dedicated Connections panel. Each connection has its own identity separate from the model ID -- so you can configure "Phi-4 (Foundry Local)" and "Phi-4 (Ollama)" as distinct connections pointing to different endpoints but using the same model.

See: https://prompty.ai/vscode/connections/