import type { HybridObject } from 'react-native-nitro-modules';
/**
 * Options for loading an embeddings model.
 */
export interface EmbeddingsLoadOptions {
    /** Callback invoked with loading progress (0-1) */
    onProgress?: (progress: number) => void;
}
/**
 * Low-level embeddings interface using MLX.
 * @internal Use the `Embeddings` export from `react-native-nitro-mlx` instead.
 */
export interface Embeddings extends HybridObject<{
    ios: 'swift';
}> {
    /**
     * Load an embedding model into memory. Downloads from HuggingFace if not already cached.
     * @param modelId - HuggingFace model ID (e.g., 'mlx-community/bge-small-en-v1.5-4bit')
     * @param options - Callback invoked with loading progress (0-1)
     */
    load(modelId: string, options?: EmbeddingsLoadOptions): Promise<void>;
    /**
     * Unload the current model and release memory.
     */
    unload(): void;
    /**
     * Embed a single text and return the vector as Float32 bytes.
     * @param text - Input text
     * @returns Float32 embedding bytes (length = `dimension` floats)
     */
    embed(text: string): Promise<ArrayBuffer>;
    /**
     * Embed a batch of texts in a single forward pass (padded to the longest).
     * @param texts - Array of input texts
     * @returns Array of Float32 embedding buffers, one per input
     */
    embedBatch(texts: string[]): Promise<ArrayBuffer[]>;
    /** Whether a model is currently loaded */
    readonly isLoaded: boolean;
    /** Output embedding dimension (e.g., 384 for bge-small) */
    readonly dimension: number;
    /** Maximum supported sequence length (tokens) */
    readonly maxSequenceLength: number;
}
//# sourceMappingURL=Embeddings.nitro.d.ts.map