import { SapChat } from './chat.ts'; import { SapEmbeddings } from './embeddings.ts'; import { SapResponses } from './responses.ts'; import type { SapOpenAiInput } from './types.ts'; /** * A pre-configured client for SAP AI Core backed by the official `openai` package. * Handles deployment resolution, authentication, and SAP-specific headers automatically. * * The `model` parameter is removed from `chat.completions.create()`, `chat.completions.parse()`, * `embeddings.create()`, and `responses.create()` signatures — SAP AI Core routes requests via * the deployment URL, so the `model` field in the request body is not used. * * Only the endpoints supported by SAP AI Core are exposed (`chat`, `embeddings`, `responses`). * Use {@link SapOpenAi.createClient} to create an instance. * @experimental This class is experimental and may change at any time without prior notice. */ export declare class SapOpenAi { /** * Creates a pre-configured {@link SapOpenAi} client for SAP AI Core. * Resolves the deployment and sets up authentication automatically. * The `model` parameter is omitted from `chat.completions.create()`, `chat.completions.parse()`, * `embeddings.create()`, and `responses.create()` — SAP AI Core routes requests via the deployment URL. * @param options - Options including model deployment, destination, API version, and client type. A plain model name string is accepted as shorthand for `{ deployment: modelName }`. * @returns A promise that resolves to a ready-to-use {@link SapOpenAi} instance. * @example * ```ts * import { SapOpenAi } from '@sap-ai-sdk/openai'; * * const client = await SapOpenAi.createClient('gpt-5.4'); * await client.chat.completions.create({ * messages: [{ role: 'user', content: 'Hello!' }] * }); * ``` */ static createClient(options: SapOpenAiInput): Promise; readonly chat: SapChat; readonly embeddings: SapEmbeddings; readonly responses: SapResponses; /** @internal — use {@link SapOpenAi.createClient} instead */ private constructor(); } //# sourceMappingURL=client.d.ts.map