crestapps-core-openai

CrestApps.Core OpenAI

The CrestApps.Core.AI.OpenAI package connects CrestApps.Core to api.openai.com and to any server that speaks the OpenAI API. The connection holds credentials and the endpoint; each deployment names a model and declares what that model can do.

Guidelines

• Register only the providers the host uses. OpenAI is the simplest first provider because it needs only an API key.
• Put the API key, endpoint and ChatApi on the connection. Put the model name and its capabilities on the deployment.
• Every configured deployment must declare its model capabilities (Features). A deployment entry with no features and no legacy Purpose/Type is skipped at startup with the warning "declares no model capabilities".
• Pick deployments through slots (AIDeploymentSlotNames) and IAIDeploymentManager.ResolveSlotAsync(...). Never hard-code a provider client in business code.
• Use crestapps-core-ai-runtime for the provider-agnostic runtime, crestapps-core-chat for completion loops, and crestapps-core-realtime-voice for voice sessions.

Registration

using CrestApps.Core.AI.OpenAI;

builder.Services.AddCrestAppsCore(crestApps => crestApps
    .AddAISuite(ai => ai
        .AddOpenAI()
    )
);

AddAISuite(...) already calls AddCoreServices(), AddCoreAIServices() and AddCoreAIOrchestration(). Without the builder, register them yourself:

builder.Services
    .AddCoreAIServices()
    .AddCoreAIOrchestration()
    .AddCoreAIOpenAI();

AddCoreAIOpenAI() registers: