@tanstack/ai 0.0.3 → 0.2.0

package/src/activities/generateTranscription/adapter.ts

@@ -0,0 +1,89 @@
+import type { TranscriptionOptions, TranscriptionResult } from '../../types'
+
+/**
+ * Configuration for transcription adapter instances
+ */
+export interface TranscriptionAdapterConfig {
+  apiKey?: string
+  baseUrl?: string
+  timeout?: number
+  maxRetries?: number
+  headers?: Record<string, string>
+}
+
+/**
+ * Transcription adapter interface with pre-resolved generics.
+ *
+ * An adapter is created by a provider function: `provider('model')` → `adapter`
+ * All type resolution happens at the provider call site, not in this interface.
+ *
+ * Generic parameters:
+ * - TModel: The specific model name (e.g., 'whisper-1')
+ * - TProviderOptions: Provider-specific options (already resolved)
+ */
+export interface TranscriptionAdapter<
+  TModel extends string = string,
+  TProviderOptions extends object = Record<string, unknown>,
+> {
+  /** Discriminator for adapter kind - used to determine API shape */
+  readonly kind: 'transcription'
+  /** Adapter name identifier */
+  readonly name: string
+  /** The model this adapter is configured for */
+  readonly model: TModel
+
+  /**
+   * @internal Type-only properties for inference. Not assigned at runtime.
+   */
+  '~types': {
+    providerOptions: TProviderOptions
+  }
+
+  /**
+   * Transcribe audio to text
+   */
+  transcribe: (
+    options: TranscriptionOptions<TProviderOptions>,
+  ) => Promise<TranscriptionResult>
+}
+
+/**
+ * A TranscriptionAdapter with any/unknown type parameters.
+ * Useful as a constraint in generic functions and interfaces.
+ */
+export type AnyTranscriptionAdapter = TranscriptionAdapter<any, any>
+
+/**
+ * Abstract base class for audio transcription adapters.
+ * Extend this class to implement a transcription adapter for a specific provider.
+ *
+ * Generic parameters match TranscriptionAdapter - all pre-resolved by the provider function.
+ */
+export abstract class BaseTranscriptionAdapter<
+  TModel extends string = string,
+  TProviderOptions extends object = Record<string, unknown>,
+> implements TranscriptionAdapter<TModel, TProviderOptions> {
+  readonly kind = 'transcription' as const
+  abstract readonly name: string
+  readonly model: TModel
+
+  // Type-only property - never assigned at runtime
+  declare '~types': {
+    providerOptions: TProviderOptions
+  }
+
+  protected config: TranscriptionAdapterConfig
+
+  constructor(config: TranscriptionAdapterConfig = {}, model: TModel) {
+    this.config = config
+    this.model = model
+  }
+
+  abstract transcribe(
+    options: TranscriptionOptions<TProviderOptions>,
+  ): Promise<TranscriptionResult>
+
+  protected generateId(): string {
+    return `${this.name}-${Date.now()}-${Math.random().toString(36).substring(7)}`
+  }
+}

package/src/activities/generateTranscription/index.ts

@@ -0,0 +1,132 @@
+/**
+ * Transcription Activity
+ *
+ * Transcribes audio to text using speech-to-text models.
+ * This is a self-contained module with implementation, types, and JSDoc.
+ */
+
+import type { TranscriptionAdapter } from './adapter'
+import type { TranscriptionResult } from '../../types'
+
+// ===========================
+// Activity Kind
+// ===========================
+
+/** The adapter kind this activity handles */
+export const kind = 'transcription' as const
+
+// ===========================
+// Type Extraction Helpers
+// ===========================
+
+/**
+ * Extract provider options from a TranscriptionAdapter via ~types.
+ */
+export type TranscriptionProviderOptions<TAdapter> =
+  TAdapter extends TranscriptionAdapter<any, any>
+    ? TAdapter['~types']['providerOptions']
+    : object
+
+// ===========================
+// Activity Options Type
+// ===========================
+
+/**
+ * Options for the transcription activity.
+ * The model is extracted from the adapter's model property.
+ *
+ * @template TAdapter - The transcription adapter type
+ */
+export interface TranscriptionActivityOptions<
+  TAdapter extends TranscriptionAdapter<string, object>,
+> {
+  /** The transcription adapter to use (must be created with a model) */
+  adapter: TAdapter & { kind: typeof kind }
+  /** The audio data to transcribe - can be base64 string, File, Blob, or Buffer */
+  audio: string | File | Blob | ArrayBuffer
+  /** The language of the audio in ISO-639-1 format (e.g., 'en') */
+  language?: string
+  /** An optional prompt to guide the transcription */
+  prompt?: string
+  /** The format of the transcription output */
+  responseFormat?: 'json' | 'text' | 'srt' | 'verbose_json' | 'vtt'
+  /** Provider-specific options for transcription */
+  modelOptions?: TranscriptionProviderOptions<TAdapter>
+}
+
+// ===========================
+// Activity Result Type
+// ===========================
+
+/** Result type for the transcription activity */
+export type TranscriptionActivityResult = Promise<TranscriptionResult>
+
+// ===========================
+// Activity Implementation
+// ===========================
+
+/**
+ * Transcription activity - converts audio to text.
+ *
+ * Uses AI speech-to-text models to transcribe audio content.
+ *
+ * @example Transcribe an audio file
+ * ```ts
+ * import { generateTranscription } from '@tanstack/ai'
+ * import { openaiTranscription } from '@tanstack/ai-openai'
+ *
+ * const result = await generateTranscription({
+ *   adapter: openaiTranscription('whisper-1'),
+ *   audio: audioFile, // File, Blob, or base64 string
+ *   language: 'en'
+ * })
+ *
+ * console.log(result.text)
+ * ```
+ *
+ * @example With verbose output for timestamps
+ * ```ts
+ * const result = await generateTranscription({
+ *   adapter: openaiTranscription('whisper-1'),
+ *   audio: audioFile,
+ *   responseFormat: 'verbose_json'
+ * })
+ *
+ * result.segments?.forEach(segment => {
+ *   console.log(`[${segment.start}s - ${segment.end}s]: ${segment.text}`)
+ * })
+ * ```
+ */
+export async function generateTranscription<
+  TAdapter extends TranscriptionAdapter<string, object>,
+>(
+  options: TranscriptionActivityOptions<TAdapter>,
+): TranscriptionActivityResult {
+  const { adapter, ...rest } = options
+  const model = adapter.model
+
+  return adapter.transcribe({ ...rest, model })
+}
+
+// ===========================
+// Options Factory
+// ===========================
+
+/**
+ * Create typed options for the generateTranscription() function without executing.
+ */
+export function createTranscriptionOptions<
+  TAdapter extends TranscriptionAdapter<string, object>,
+>(
+  options: TranscriptionActivityOptions<TAdapter>,
+): TranscriptionActivityOptions<TAdapter> {
+  return options
+}
+
+// Re-export adapter types
+export type {
+  TranscriptionAdapter,
+  TranscriptionAdapterConfig,
+  AnyTranscriptionAdapter,
+} from './adapter'
+export { BaseTranscriptionAdapter } from './adapter'

package/src/activities/generateVideo/adapter.ts

@@ -0,0 +1,116 @@
+import type {
+  VideoGenerationOptions,
+  VideoJobResult,
+  VideoStatusResult,
+  VideoUrlResult,
+} from '../../types'
+
+/**
+ * Configuration for video adapter instances
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ */
+export interface VideoAdapterConfig {
+  apiKey?: string
+  baseUrl?: string
+  timeout?: number
+  maxRetries?: number
+  headers?: Record<string, string>
+}
+
+/**
+ * Video adapter interface with pre-resolved generics.
+ *
+ * An adapter is created by a provider function: `provider('model')` → `adapter`
+ * All type resolution happens at the provider call site, not in this interface.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ *
+ * Generic parameters:
+ * - TModel: The specific model name (e.g., 'sora-2')
+ * - TProviderOptions: Provider-specific options (already resolved)
+ */
+export interface VideoAdapter<
+  TModel extends string = string,
+  TProviderOptions extends object = Record<string, unknown>,
+> {
+  /** Discriminator for adapter kind - used to determine API shape */
+  readonly kind: 'video'
+  /** Adapter name identifier */
+  readonly name: string
+  /** The model this adapter is configured for */
+  readonly model: TModel
+
+  /**
+   * @internal Type-only properties for inference. Not assigned at runtime.
+   */
+  '~types': {
+    providerOptions: TProviderOptions
+  }
+
+  /**
+   * Create a new video generation job.
+   * Returns a job ID that can be used to poll for status and retrieve the video.
+   */
+  createVideoJob: (
+    options: VideoGenerationOptions<TProviderOptions>,
+  ) => Promise<VideoJobResult>
+
+  /**
+   * Get the current status of a video generation job.
+   */
+  getVideoStatus: (jobId: string) => Promise<VideoStatusResult>
+
+  /**
+   * Get the URL to download/view the generated video.
+   * Should only be called after status is 'completed'.
+   */
+  getVideoUrl: (jobId: string) => Promise<VideoUrlResult>
+}
+
+/**
+ * A VideoAdapter with any/unknown type parameters.
+ * Useful as a constraint in generic functions and interfaces.
+ */
+export type AnyVideoAdapter = VideoAdapter<any, any>
+
+/**
+ * Abstract base class for video generation adapters.
+ * Extend this class to implement a video adapter for a specific provider.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ *
+ * Generic parameters match VideoAdapter - all pre-resolved by the provider function.
+ */
+export abstract class BaseVideoAdapter<
+  TModel extends string = string,
+  TProviderOptions extends object = Record<string, unknown>,
+> implements VideoAdapter<TModel, TProviderOptions> {
+  readonly kind = 'video' as const
+  abstract readonly name: string
+  readonly model: TModel
+
+  // Type-only property - never assigned at runtime
+  declare '~types': {
+    providerOptions: TProviderOptions
+  }
+
+  protected config: VideoAdapterConfig
+
+  constructor(config: VideoAdapterConfig = {}, model: TModel) {
+    this.config = config
+    this.model = model
+  }
+
+  abstract createVideoJob(
+    options: VideoGenerationOptions<TProviderOptions>,
+  ): Promise<VideoJobResult>
+
+  abstract getVideoStatus(jobId: string): Promise<VideoStatusResult>
+
+  abstract getVideoUrl(jobId: string): Promise<VideoUrlResult>
+
+  protected generateId(): string {
+    return `${this.name}-${Date.now()}-${Math.random().toString(36).substring(7)}`
+  }
+}

package/src/activities/generateVideo/index.ts

@@ -0,0 +1,261 @@
+/**
+ * Video Activity (Experimental)
+ *
+ * Generates videos from text prompts using a jobs/polling architecture.
+ * This is a self-contained module with implementation, types, and JSDoc.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ */
+
+import type { VideoAdapter } from './adapter'
+import type {
+  VideoJobResult,
+  VideoStatusResult,
+  VideoUrlResult,
+} from '../../types'
+
+// ===========================
+// Activity Kind
+// ===========================
+
+/** The adapter kind this activity handles */
+export const kind = 'video' as const
+
+// ===========================
+// Type Extraction Helpers
+// ===========================
+
+/**
+ * Extract provider options from a VideoAdapter via ~types.
+ */
+export type VideoProviderOptions<TAdapter> =
+  TAdapter extends VideoAdapter<any, any>
+    ? TAdapter['~types']['providerOptions']
+    : object
+
+// ===========================
+// Activity Options Types
+// ===========================
+
+/**
+ * Base options shared by all video activity operations.
+ * The model is extracted from the adapter's model property.
+ */
+interface VideoActivityBaseOptions<
+  TAdapter extends VideoAdapter<string, object>,
+> {
+  /** The video adapter to use (must be created with a model) */
+  adapter: TAdapter & { kind: typeof kind }
+}
+
+/**
+ * Options for creating a new video generation job.
+ * The model is extracted from the adapter's model property.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ */
+export interface VideoCreateOptions<
+  TAdapter extends VideoAdapter<string, object>,
+> extends VideoActivityBaseOptions<TAdapter> {
+  /** Request type - create a new job (default if not specified) */
+  request?: 'create'
+  /** Text description of the desired video */
+  prompt: string
+  /** Video size in WIDTHxHEIGHT format (e.g., "1280x720") */
+  size?: string
+  /** Video duration in seconds */
+  duration?: number
+  /** Provider-specific options for video generation */
+  modelOptions?: VideoProviderOptions<TAdapter>
+}
+
+/**
+ * Options for polling the status of a video generation job.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ */
+export interface VideoStatusOptions<
+  TAdapter extends VideoAdapter<string, object>,
+> extends VideoActivityBaseOptions<TAdapter> {
+  /** Request type - get job status */
+  request: 'status'
+  /** The job ID to check status for */
+  jobId: string
+}
+
+/**
+ * Options for getting the URL of a completed video.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ */
+export interface VideoUrlOptions<
+  TAdapter extends VideoAdapter<string, object>,
+> extends VideoActivityBaseOptions<TAdapter> {
+  /** Request type - get video URL */
+  request: 'url'
+  /** The job ID to get URL for */
+  jobId: string
+}
+
+/**
+ * Union type for all video activity options.
+ * Discriminated by the `request` field.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ */
+export type VideoActivityOptions<
+  TAdapter extends VideoAdapter<string, object>,
+  TRequest extends 'create' | 'status' | 'url' = 'create',
+> = TRequest extends 'status'
+  ? VideoStatusOptions<TAdapter>
+  : TRequest extends 'url'
+    ? VideoUrlOptions<TAdapter>
+    : VideoCreateOptions<TAdapter>
+
+// ===========================
+// Activity Result Types
+// ===========================
+
+/**
+ * Result type for the video activity, based on request type.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ */
+export type VideoActivityResult<
+  TRequest extends 'create' | 'status' | 'url' = 'create',
+> = TRequest extends 'status'
+  ? Promise<VideoStatusResult>
+  : TRequest extends 'url'
+    ? Promise<VideoUrlResult>
+    : Promise<VideoJobResult>
+
+// ===========================
+// Activity Implementation
+// ===========================
+
+/**
+ * Generate video - creates a video generation job from a text prompt.
+ *
+ * Uses AI video generation models to create videos based on natural language descriptions.
+ * Unlike image generation, video generation is asynchronous and requires polling for completion.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ *
+ * @example Create a video generation job
+ * ```ts
+ * import { generateVideo } from '@tanstack/ai'
+ * import { openaiVideo } from '@tanstack/ai-openai'
+ *
+ * // Start a video generation job
+ * const { jobId } = await generateVideo({
+ *   adapter: openaiVideo('sora-2'),
+ *   prompt: 'A cat chasing a dog in a sunny park'
+ * })
+ *
+ * console.log('Job started:', jobId)
+ * ```
+ */
+export async function generateVideo<
+  TAdapter extends VideoAdapter<string, object>,
+>(options: VideoCreateOptions<TAdapter>): Promise<VideoJobResult> {
+  const { adapter, prompt, size, duration, modelOptions } = options
+  const model = adapter.model
+
+  return adapter.createVideoJob({
+    model,
+    prompt,
+    size,
+    duration,
+    modelOptions,
+  })
+}
+
+/**
+ * Get video job status - returns the current status, progress, and URL if available.
+ *
+ * This function combines status checking and URL retrieval. If the job is completed,
+ * it will automatically fetch and include the video URL.
+ *
+ * @experimental Video generation is an experimental feature and may change.
+ *
+ * @example Check job status
+ * ```ts
+ * import { getVideoJobStatus } from '@tanstack/ai'
+ * import { openaiVideo } from '@tanstack/ai-openai'
+ *
+ * const result = await getVideoJobStatus({
+ *   adapter: openaiVideo('sora-2'),
+ *   jobId: 'job-123'
+ * })
+ *
+ * console.log('Status:', result.status)
+ * console.log('Progress:', result.progress)
+ * if (result.url) {
+ *   console.log('Video URL:', result.url)
+ * }
+ * ```
+ */
+export async function getVideoJobStatus<
+  TAdapter extends VideoAdapter<string, object>,
+>(options: {
+  adapter: TAdapter & { kind: typeof kind }
+  jobId: string
+}): Promise<{
+  status: 'pending' | 'processing' | 'completed' | 'failed'
+  progress?: number
+  url?: string
+  error?: string
+}> {
+  const { adapter, jobId } = options
+
+  // Get status first
+  const statusResult = await adapter.getVideoStatus(jobId)
+
+  // If completed, also get the URL
+  if (statusResult.status === 'completed') {
+    try {
+      const urlResult = await adapter.getVideoUrl(jobId)
+      return {
+        status: statusResult.status,
+        progress: statusResult.progress,
+        url: urlResult.url,
+      }
+    } catch (error) {
+      // If URL fetch fails, still return status
+      return {
+        status: statusResult.status,
+        progress: statusResult.progress,
+        error:
+          error instanceof Error ? error.message : 'Failed to get video URL',
+      }
+    }
+  }
+
+  // Return status for non-completed jobs
+  return {
+    status: statusResult.status,
+    progress: statusResult.progress,
+    error: statusResult.error,
+  }
+}
+
+// ===========================
+// Options Factory
+// ===========================
+
+/**
+ * Create typed options for the generateVideo() function without executing.
+ */
+export function createVideoOptions<
+  TAdapter extends VideoAdapter<string, object>,
+>(options: VideoCreateOptions<TAdapter>): VideoCreateOptions<TAdapter> {
+  return options
+}
+
+// Re-export adapter types
+export type {
+  VideoAdapter,
+  VideoAdapterConfig,
+  AnyVideoAdapter,
+} from './adapter'
+export { BaseVideoAdapter } from './adapter'