Interface AIProvider

interface AIProvider {
    providerId: string;
    providerName: string;
    supportedModels: string[];
    discoverModels(): Promise<string[]>;
    isModelSupported(modelId: string): boolean;
    process(request: AIRequest): Promise<AIResponse>;
    processStream?(
        request: AIRequest,
    ): AsyncGenerator<AIStreamChunk, void, void>;
    testConnection(): Promise<boolean>;
}

Implemented by

BaseProvider

Index

Properties

providerId providerName supportedModels

Methods

discoverModels isModelSupported process processStream? testConnection

Properties

`Readonly`providerId

providerId: string

`Readonly`providerName

providerName: string

`Readonly`supportedModels

supportedModels: string[]

Methods

discoverModels

discoverModels(): Promise<string[]>
Returns Promise<string[]>
- Defined in types/index.ts:25

isModelSupported

isModelSupported(modelId: string): boolean
Parameters
- modelId: string
Returns boolean
- Defined in types/index.ts:23

process

process(request: AIRequest): Promise<AIResponse>
Parameters
- request: AIRequest
Returns Promise<AIResponse>
- Defined in types/index.ts:14

`Optional`processStream

processStream?(request: AIRequest): AsyncGenerator<AIStreamChunk, void, void>
The same completion, delivered as it is written.

Optional on the interface so a provider written before streaming existed still satisfies it; BaseProvider supplies a one-chunk implementation, so every provider that extends it can be consumed as a stream regardless.
Parameters
- request: AIRequest
Returns AsyncGenerator<AIStreamChunk, void, void>
- Defined in types/index.ts:22

testConnection

testConnection(): Promise<boolean>
Returns Promise<boolean>
- Defined in types/index.ts:24