Public API

The strongest public surface in AI Providers is the service plus provider-asset abstraction. Editor panels and inspectors matter for usability, but the reusable runtime contract is the real API core.

Main reusable APIs

API	Purpose	Practical stability
`AiProvidersService`	Shared prompt transport and normalized request execution.	High
`LlmProviderDefinitionBase`	Base class for all provider assets.	High
`LlmChatRequest`	Chat payload with messages, temperature, max tokens, and model override.	High
`LlmMessage`	Role/content chat message item.	High
`AiLocalSettings`	Project-scoped local provider key storage helper.	Medium

DTO surface

public class LlmChatRequest
{
    public List<LlmMessage> messages = new List<LlmMessage>();
    public float temperature = 0.7f;
    public int maxTokens = 512;
    public string modelOverride;
}

Built-in provider types

The product ships concrete provider asset types for OpenAI, DeepSeek, Gemini, Ollama, and Custom JSON. Those are the public provider families to document, rather than hidden internal editor helper classes.