egaki 0.5.0 → 0.7.0

package/src/cli/generate.ts

@@ -0,0 +1,912 @@
+// Programmatic API for egaki image and video generation.
+// Returns Error | Result (errore style) instead of process.exit/console.error.
+//
+// Two variants for each function:
+//   - `generateImage(opts)` — cached, writes to public/generated/, returns { src: string }
+//   - `generateImageUncached(opts)` — raw, returns bytes in GenerateImageResult
+//
+// The CLI uses *Uncached variants. The cached variants are for the Vite MDX
+// video framework and .server.tsx components.
+//
+// Usage:
+//   import { generateImage, generateImageUncached } from 'egaki/generate'
+//   const cached = await generateImage({ prompt: 'a sunset' })
+//   if (!(cached instanceof Error)) cached.src // '/generated/image/sunset-a1b2c3d4.png'
+//   const raw = await generateImageUncached({ prompt: 'a sunset' })
+//   if (!(raw instanceof Error)) raw.images[0].uint8Array // raw bytes
+import {
+  generateImage as aiGenerateImage,
+  generateText,
+  experimental_generateVideo as aiGenerateVideo,
+} from 'ai'
+import type { XaiImageModelOptions, XaiVideoModelOptions } from '@ai-sdk/xai'
+import type { GoogleImageModelOptions, GoogleLanguageModelOptions, GoogleVideoModelOptions } from '@ai-sdk/google'
+import type { GoogleVertexImageModelOptions, GoogleVertexVideoModelOptions } from '@ai-sdk/google-vertex'
+import type { FalImageModelOptions, FalVideoModelOptions } from '@ai-sdk/fal'
+import type { ByteDanceVideoProviderOptions } from '@ai-sdk/bytedance'
+import { createParser, type EventSourceMessage } from 'eventsource-parser'
+import {
+  injectCredentialsToEnv,
+  getChatGptAuth,
+  saveChatGptAuth,
+} from './credentials.js'
+import {
+  getModelConfig,
+  createImageModel,
+  createTextModel,
+  createVideoModel,
+  shouldUseResponsesApi,
+  DEFAULT_MODEL,
+  DEFAULT_VIDEO_MODEL,
+  ValidationError,
+} from './models.js'
+
+export { ValidationError }
+import { getValidChatGptAuth } from './chatgpt-auth.js'
+
+// ─── autocomplete-friendly union types ───────────────────────────────────────
+// `(string & {})` lets users pass arbitrary strings (new models, custom
+// endpoints) while still providing autocomplete for known values.
+
+/** Image model IDs from the catalog. Accepts arbitrary strings too. */
+export type ImageModelId =
+  | 'imagen-4.0-generate-001' | 'imagen-4.0-ultra-generate-001' | 'imagen-4.0-fast-generate-001'
+  | 'gemini-2.0-flash-exp-image-generation' | 'gemini-2.5-flash-image' | 'gemini-3-pro-image-preview' | 'gemini-3.1-flash-image-preview'
+  | 'vertex/imagen-4.0-generate-001' | 'vertex/imagen-4.0-ultra-generate-001' | 'vertex/imagen-4.0-fast-generate-001'
+  | 'vertex/gemini-2.5-flash-image' | 'vertex/gemini-3-pro-image-preview' | 'vertex/gemini-3.1-flash-image-preview'
+  | 'flux-kontext-max' | 'flux-kontext-pro' | 'flux-pro-1.0-fill' | 'flux-pro-1.1' | 'flux-pro-1.1-ultra'
+  | 'flux-2-pro' | 'flux-2-max' | 'flux-2-flex' | 'flux-2-klein-9b' | 'flux-2-klein-4b'
+  | 'recraft-v2' | 'recraft-v3' | 'recraft-v4' | 'recraft-v4-pro' | 'recraft-v4.1' | 'recraft-v4.1-pro' | 'recraft-v4.1-utility' | 'recraft-v4.1-utility-pro'
+  | 'seedream-5.0-lite' | 'seedream-4.5' | 'seedream-4.0'
+  | 'grok-imagine-image' | 'grok-imagine-image-quality'
+  | 'dall-e-2' | 'dall-e-3' | 'gpt-image-1' | 'gpt-image-1-mini' | 'gpt-image-1.5' | 'gpt-image-2' | 'chatgpt-image-latest'
+  | (string & {})
+
+/** Video model IDs from the catalog. Accepts arbitrary strings too. */
+export type VideoModelId =
+  | 'veo-3.1-generate-001' | 'veo-3.1-fast-generate-001' | 'veo-3.0-generate-001' | 'veo-3.0-fast-generate-001'
+  | 'vertex/veo-3.1-generate-001' | 'vertex/veo-3.1-fast-generate-001'
+  | 'grok-imagine-video' | 'grok-imagine-video-1.5'
+  | 'kling-v3.0-t2v' | 'kling-v3.0-i2v' | 'kling-v3.0-motion-control' | 'kling-v2.6-t2v' | 'kling-v2.6-i2v' | 'kling-v2.6-motion-control'
+  | 'kling-v2.5-turbo-t2v' | 'kling-v2.5-turbo-i2v'
+  | 'wan-v2.6-t2v' | 'wan-v2.6-i2v' | 'wan-v2.6-i2v-flash' | 'wan-v2.6-r2v' | 'wan-v2.6-r2v-flash' | 'wan-v2.5-t2v-preview'
+  | 'seedance-2.0' | 'seedance-2.0-fast' | 'seedance-v1.5-pro' | 'seedance-v1.0-pro' | 'seedance-v1.0-pro-fast' | 'seedance-v1.0-lite-t2v' | 'seedance-v1.0-lite-i2v'
+  | 'luma-ray-2' | 'minimax-video' | 'hunyuan-video'
+  | (string & {})
+
+/** Common aspect ratios across providers. */
+export type AspectRatio =
+  | '1:1' | '3:4' | '4:3' | '9:16' | '16:9' | '2:3' | '3:2' | '4:5' | '5:4' | '9:21' | '21:9'
+  | (string & {})
+
+/** Image quality presets (varies by provider). */
+export type ImageQuality = 'auto' | 'low' | 'medium' | 'high' | 'standard' | (string & {})
+
+/** Image output formats. */
+export type ImageOutputFormat = 'png' | 'jpeg' | 'webp' | (string & {})
+
+// ─── public types ────────────────────────────────────────────────────────────
+
+export type GeneratedFile = {
+  uint8Array: Uint8Array
+  mediaType: string
+}
+
+export interface GenerateImageOptions {
+  prompt: string
+  /** Model ID. Defaults to DEFAULT_MODEL if omitted. */
+  model?: ImageModelId
+  count?: number
+  aspectRatio?: AspectRatio
+  seed?: number
+  inputImages?: Uint8Array[]
+  maskImage?: Uint8Array
+  allowPeople?: boolean
+  quality?: ImageQuality
+  resolution?: string
+  outputFormat?: ImageOutputFormat
+  negativePrompt?: string
+  /** Only for text-model image generation (Gemini). */
+  imageSize?: '1K' | '2K' | '4K'
+}
+
+export interface GenerateImageResult {
+  images: GeneratedFile[]
+  model: string
+  cost: number | null
+  usage?: { inputTokens?: number; outputTokens?: number }
+  warnings?: unknown[]
+  /** Text response from text-model path (Gemini). */
+  text?: string
+  /** Revised prompt from ChatGPT Responses API path. */
+  revisedPrompt?: string | null
+}
+
+export interface GenerateVideoOptions {
+  prompt: string
+  /** Model ID. Defaults to DEFAULT_VIDEO_MODEL if omitted. */
+  model?: VideoModelId
+  count?: number
+  aspectRatio?: AspectRatio
+  resolution?: string
+  duration?: number
+  fps?: number
+  seed?: number
+  inputImage?: Uint8Array
+  mode?: 'edit-video' | 'extend-video' | 'reference-to-video'
+  /** URL for video editing/extension input. Must be a URL (not a local path). */
+  videoUrl?: string
+  /** Reference image URLs for r2v generation. Must be URLs (not local paths). */
+  referenceImages?: string[]
+  negativePrompt?: string
+}
+
+export interface GenerateVideoResult {
+  videos: GeneratedFile[]
+  model: string
+  cost: number | null
+  warnings?: unknown[]
+  /** Raw responses from the AI SDK (provider-specific). */
+  responses?: unknown[]
+}
+
+// ─── OpenAI provider options type (not exported by @ai-sdk/openai) ───────────
+
+type OpenAIImageProviderOptions = {
+  quality?: 'standard' | 'low' | 'medium' | 'high' | 'auto'
+  outputFormat?: 'png' | 'jpeg' | 'webp'
+  outputCompression?: number
+  size?: `${number}x${number}`
+  partialImages?: number | null
+  background?: 'auto' | 'opaque' | 'transparent'
+}
+
+// ─── type-safe provider option builders ──────────────────────────────────────
+
+type ProviderOptionsResult = Record<string, Record<string, string | number | boolean | string[] | null | undefined | Record<string, string | number | boolean | null | undefined>>>
+
+export function buildImageProviderOptions(
+  provider: string,
+  opts: {
+    aspectRatio?: `${number}:${number}`
+    allowPeople: boolean
+    quality?: string
+    resolution?: string
+    outputFormat?: string
+    negativePrompt?: string
+  },
+): ProviderOptionsResult {
+  switch (provider) {
+    case 'google': {
+      const googleOpts = {
+        ...(opts.allowPeople ? { personGeneration: 'allow_all' as const } : {}),
+        ...(opts.aspectRatio ? { aspectRatio: opts.aspectRatio as GoogleImageModelOptions['aspectRatio'] } : {}),
+      } satisfies GoogleImageModelOptions
+      return { google: googleOpts }
+    }
+    case 'vertex': {
+      const vertexOpts = {
+        ...(opts.allowPeople ? { personGeneration: 'allow_all' as const } : {}),
+        ...(opts.negativePrompt ? { negativePrompt: opts.negativePrompt } : {}),
+      } satisfies GoogleVertexImageModelOptions
+      return { vertex: vertexOpts }
+    }
+    case 'xai': {
+      const xaiOpts = {
+        ...(opts.quality ? { quality: opts.quality as XaiImageModelOptions['quality'] } : {}),
+        ...(opts.resolution ? { resolution: opts.resolution as XaiImageModelOptions['resolution'] } : {}),
+        ...(opts.outputFormat ? { output_format: opts.outputFormat } : {}),
+      } satisfies XaiImageModelOptions
+      return { xai: xaiOpts }
+    }
+    case 'openai': {
+      const openaiOpts = {
+        ...(opts.quality ? { quality: opts.quality as OpenAIImageProviderOptions['quality'] } : {}),
+        ...(opts.outputFormat ? { outputFormat: opts.outputFormat as OpenAIImageProviderOptions['outputFormat'] } : {}),
+      } satisfies OpenAIImageProviderOptions
+      return { openai: openaiOpts }
+    }
+    case 'fal': {
+      const falOpts = {
+        ...(opts.outputFormat ? { outputFormat: opts.outputFormat } : {}),
+        ...(opts.negativePrompt ? { negativePrompt: opts.negativePrompt } : {}),
+      } satisfies FalImageModelOptions
+      return { fal: falOpts }
+    }
+    default:
+      return {}
+  }
+}
+
+export function buildVideoProviderOptions(
+  provider: string,
+  opts: {
+    resolution?: string
+    mode?: string
+    videoUrl?: string
+    referenceImages?: string[]
+    negativePrompt?: string
+    model: string
+  },
+): ProviderOptionsResult | undefined {
+  switch (provider) {
+    case 'xai': {
+      const base = {
+        ...(opts.resolution ? { resolution: opts.resolution as '480p' | '720p' } : {}),
+      }
+      if (opts.mode === 'edit-video' && opts.videoUrl) {
+        const xaiOpts = { ...base, mode: 'edit-video' as const, videoUrl: opts.videoUrl } satisfies XaiVideoModelOptions
+        return { xai: xaiOpts }
+      }
+      if (opts.mode === 'extend-video' && opts.videoUrl) {
+        const xaiOpts = { ...base, mode: 'extend-video' as const, videoUrl: opts.videoUrl } satisfies XaiVideoModelOptions
+        return { xai: xaiOpts }
+      }
+      if (opts.mode === 'reference-to-video' && opts.referenceImages) {
+        const xaiOpts = { ...base, mode: 'reference-to-video' as const, referenceImageUrls: opts.referenceImages } satisfies XaiVideoModelOptions
+        return { xai: xaiOpts }
+      }
+      if (Object.keys(base).length === 0) return undefined
+      const xaiOpts = { ...base } satisfies XaiVideoModelOptions
+      return { xai: xaiOpts }
+    }
+    case 'bytedance': {
+      const bytedanceOpts = {
+        ...(opts.resolution ? { resolution: opts.resolution } : {}),
+        ...(opts.referenceImages ? { referenceImages: opts.referenceImages } : {}),
+      } satisfies ByteDanceVideoProviderOptions
+      if (Object.keys(bytedanceOpts).length === 0) return undefined
+      return { bytedance: bytedanceOpts }
+    }
+    case 'fal': {
+      const falOpts = {
+        ...(opts.resolution ? { resolution: opts.resolution } : {}),
+        ...(opts.negativePrompt ? { negativePrompt: opts.negativePrompt } : {}),
+      } satisfies FalVideoModelOptions
+      if (Object.keys(falOpts).length === 0) return undefined
+      return { fal: falOpts }
+    }
+    case 'google': {
+      const googleOpts = {
+        ...(opts.negativePrompt ? { negativePrompt: opts.negativePrompt } : {}),
+      } satisfies GoogleVideoModelOptions
+      if (Object.keys(googleOpts).length === 0) return undefined
+      return { google: googleOpts }
+    }
+    case 'vertex': {
+      const vertexOpts = {
+        ...(opts.negativePrompt ? { negativePrompt: opts.negativePrompt } : {}),
+      } satisfies GoogleVertexVideoModelOptions
+      if (Object.keys(vertexOpts).length === 0) return undefined
+      return { vertex: vertexOpts }
+    }
+    default:
+      return undefined
+  }
+}
+
+// ─── cost calculation ────────────────────────────────────────────────────────
+
+export function calculateCost(
+  cost: {
+    type: 'per-image'
+    perImage: number
+  } | {
+    type: 'per-token'
+    inputPerM: number
+    outputPerM: number
+  } | {
+    type: 'per-video-second'
+    defaultDurationSec: number
+    tiers: Array<{ resolution?: string; costPerSecond: number; hasVideoInput?: boolean }>
+  } | {
+    type: 'unknown'
+  },
+  usage: {
+    inputTokens?: number
+    outputTokens?: number
+    imagesGenerated?: number
+    videosGenerated?: number
+    durationSeconds?: number
+    resolution?: string
+    hasVideoInput?: boolean
+  },
+  count: number = 1,
+): number | null {
+  if (cost.type === 'per-image') {
+    return cost.perImage * count
+  }
+  if (cost.type === 'per-token' && usage.inputTokens != null && usage.outputTokens != null) {
+    return (
+      (usage.inputTokens * cost.inputPerM + usage.outputTokens * cost.outputPerM) / 1_000_000
+    )
+  }
+  if (cost.type === 'per-video-second') {
+    const durationSec = usage.durationSeconds ?? cost.defaultDurationSec
+    const resolution = normalizeResolutionKey(usage.resolution)
+    const hasVideoInput = usage.hasVideoInput ?? false
+    const tier =
+      cost.tiers.find((t) =>
+        normalizeResolutionKey(t.resolution) === resolution &&
+        (t.hasVideoInput == null || t.hasVideoInput === hasVideoInput),
+      ) ??
+      cost.tiers.find((t) => normalizeResolutionKey(t.resolution) === resolution) ??
+      cost.tiers[0]
+    if (!tier) {
+      return null
+    }
+    return tier.costPerSecond * durationSec * count
+  }
+  return null
+}
+
+function normalizeResolutionKey(input?: string): string | undefined {
+  if (!input) return undefined
+  const normalized = input.trim().toLowerCase()
+  if (normalized === '1920x1080') return '1080p'
+  if (normalized === '1280x720') return '720p'
+  if (normalized === '854x480' || normalized === '848x480') return '480p'
+  return normalized
+}
+
+// ─── image byte helpers ──────────────────────────────────────────────────────
+
+function inferImageMediaType(image: Uint8Array): string {
+  if (image.length >= 8 && image[0] === 0x89 && image[1] === 0x50 && image[2] === 0x4e && image[3] === 0x47) {
+    return 'image/png'
+  }
+  if (image.length >= 3 && image[0] === 0xff && image[1] === 0xd8 && image[2] === 0xff) {
+    return 'image/jpeg'
+  }
+  if (
+    image.length >= 12 &&
+    image[0] === 0x52 && image[1] === 0x49 && image[2] === 0x46 && image[3] === 0x46 &&
+    image[8] === 0x57 && image[9] === 0x45 && image[10] === 0x42 && image[11] === 0x50
+  ) {
+    return 'image/webp'
+  }
+  if (
+    image.length >= 6 &&
+    image[0] === 0x47 && image[1] === 0x49 && image[2] === 0x46 && image[3] === 0x38
+  ) {
+    return 'image/gif'
+  }
+  return 'image/png'
+}
+
+function imageBytesToDataUrl(image: Uint8Array): string {
+  const mediaType = inferImageMediaType(image)
+  return `data:${mediaType};base64,${Buffer.from(image).toString('base64')}`
+}
+
+function chatGptImageSizeFromAspectRatio(
+  aspectRatio: `${number}:${number}`,
+): '1024x1024' | '1536x1024' | '1024x1536' | undefined {
+  switch (aspectRatio) {
+    case '1:1':
+      return '1024x1024'
+    case '3:2':
+      return '1536x1024'
+    case '2:3':
+      return '1024x1536'
+    default:
+      return undefined
+  }
+}
+
+function isAspectRatio(input: string): input is `${number}:${number}` {
+  return /^\d+:\d+$/.test(input)
+}
+
+/** Convert any thrown value into an Error instance. */
+function toError(err: unknown): Error {
+  if (err instanceof Error) return err
+  return new Error(String(err))
+}
+
+// ─── generateImage ───────────────────────────────────────────────────────────
+
+/**
+ * Generate images from a text prompt. Auto-detects which generation strategy
+ * to use (image API, text model, or ChatGPT Responses API) based on model ID.
+ * Shares credentials and subscription with the CLI.
+ */
+export async function generateImageUncached(opts: GenerateImageOptions): Promise<Error | GenerateImageResult> {
+  injectCredentialsToEnv()
+
+  const model = opts.model ?? DEFAULT_MODEL
+  const count = opts.count ?? 1
+  const config = getModelConfig(model)
+  if (config instanceof Error) return config
+
+  const parsedAspectRatio = opts.aspectRatio && isAspectRatio(opts.aspectRatio)
+    ? opts.aspectRatio
+    : undefined
+  const inputImages = opts.inputImages ?? []
+
+  // ChatGPT OAuth + OpenAI image models → Responses API path
+  const useResponsesApi = config.strategy === 'image' && shouldUseResponsesApi(model)
+
+  if (useResponsesApi) {
+    return generateWithResponsesApi({
+      prompt: opts.prompt,
+      model,
+      count,
+      aspectRatio: parsedAspectRatio,
+      seed: opts.seed,
+      maskImage: opts.maskImage,
+      inputImages,
+    })
+  }
+
+  if (config.strategy === 'image') {
+    return generateWithImageModel({
+      prompt: opts.prompt,
+      model,
+      count,
+      aspectRatio: parsedAspectRatio,
+      seed: opts.seed,
+      inputImages,
+      maskImage: opts.maskImage,
+      allowPeople: opts.allowPeople ?? false,
+      quality: opts.quality,
+      resolution: opts.resolution,
+      outputFormat: opts.outputFormat,
+      negativePrompt: opts.negativePrompt,
+    })
+  }
+
+  // Text model path (Gemini multimodal)
+  return generateWithTextModelPath({
+    prompt: opts.prompt,
+    model,
+    inputImages,
+    imageSize: opts.imageSize,
+    aspectRatio: opts.aspectRatio,
+  })
+}
+
+// ─── generateVideo ───────────────────────────────────────────────────────────
+
+/**
+ * Generate videos from a text prompt (or image+text for i2v models).
+ * Shares credentials and subscription with the CLI.
+ */
+export async function generateVideoUncached(opts: GenerateVideoOptions): Promise<Error | GenerateVideoResult> {
+  injectCredentialsToEnv()
+
+  const model = opts.model ?? DEFAULT_VIDEO_MODEL
+  const count = opts.count ?? 1
+  const config = getModelConfig(model)
+  if (config instanceof Error) return config
+
+  if (config.strategy !== 'video') {
+    return new ValidationError(`Model ${model} is not a video model`)
+  }
+
+  const isVideoInputMode = opts.mode === 'edit-video' || opts.mode === 'extend-video'
+
+  // Validate mode-specific required inputs
+  if (isVideoInputMode && !opts.videoUrl) {
+    return new ValidationError(`videoUrl is required with mode ${opts.mode}`)
+  }
+  if (opts.mode === 'reference-to-video' && (!opts.referenceImages || opts.referenceImages.length === 0)) {
+    return new ValidationError('referenceImages is required with mode reference-to-video')
+  }
+  if (opts.referenceImages?.length && !config.providerOptions?.some((opt) => opt.flag === 'reference-images')) {
+    return new ValidationError(`Model ${model} does not support referenceImages`)
+  }
+  if (opts.videoUrl && !config.providerOptions?.some((opt) => opt.flag === 'video-url')) {
+    return new ValidationError(`Model ${model} does not support video URL input for editing/extension`)
+  }
+
+  const videoModel = await createVideoModel(model)
+  if (videoModel instanceof Error) return videoModel
+
+  const providerOptions = buildVideoProviderOptions(config.provider, {
+    resolution: opts.resolution,
+    mode: opts.mode,
+    videoUrl: opts.videoUrl,
+    referenceImages: opts.referenceImages,
+    negativePrompt: opts.negativePrompt,
+    model,
+  })
+
+  let result
+  try {
+    result = await aiGenerateVideo({
+      model: videoModel,
+      prompt: opts.inputImage
+        ? { image: opts.inputImage, text: opts.prompt }
+        : opts.prompt,
+      n: count,
+      ...(opts.aspectRatio ? { aspectRatio: opts.aspectRatio as `${number}:${number}` } : {}),
+      ...(opts.resolution ? { resolution: opts.resolution as `${number}x${number}` } : {}),
+      ...(opts.duration != null ? { duration: opts.duration } : {}),
+      ...(opts.fps != null ? { fps: opts.fps } : {}),
+      ...(opts.seed != null ? { seed: opts.seed } : {}),
+      ...(providerOptions ? { providerOptions } : {}),
+    })
+  } catch (err) {
+    return toError(err)
+  }
+
+  const videos = result.videos.map((v: { uint8Array: Uint8Array; mediaType: string }) => ({
+    uint8Array: v.uint8Array,
+    mediaType: v.mediaType,
+  }))
+
+  const cost = calculateCost(config.cost, {
+    videosGenerated: result.videos.length,
+    durationSeconds: opts.duration,
+    resolution: opts.resolution,
+    hasVideoInput: Boolean(opts.inputImage || opts.videoUrl || opts.referenceImages?.length),
+  }, result.videos.length)
+
+  return {
+    videos,
+    model,
+    cost,
+    warnings: result.warnings,
+    responses: result.responses,
+  }
+}
+
+// ─── internal generation paths ───────────────────────────────────────────────
+
+async function generateWithImageModel(opts: {
+  prompt: string
+  model: string
+  count: number
+  aspectRatio?: `${number}:${number}`
+  seed?: number
+  inputImages: Uint8Array[]
+  maskImage?: Uint8Array
+  allowPeople: boolean
+  quality?: string
+  resolution?: string
+  outputFormat?: string
+  negativePrompt?: string
+}): Promise<Error | GenerateImageResult> {
+  const imagePrompt = opts.inputImages.length > 0
+    ? { text: opts.prompt, images: opts.inputImages, ...(opts.maskImage ? { mask: opts.maskImage } : {}) }
+    : opts.prompt
+
+  const config = getModelConfig(opts.model)
+  if (config instanceof Error) return config
+
+  const imageModel = await createImageModel(opts.model)
+  if (imageModel instanceof Error) return imageModel
+
+  const providerOptions = buildImageProviderOptions(config.provider, {
+    aspectRatio: opts.aspectRatio,
+    allowPeople: opts.allowPeople,
+    quality: opts.quality,
+    resolution: opts.resolution,
+    outputFormat: opts.outputFormat,
+    negativePrompt: opts.negativePrompt,
+  })
+
+  let result
+  try {
+    result = await aiGenerateImage({
+      model: imageModel,
+      prompt: imagePrompt,
+      n: opts.count,
+      ...(opts.aspectRatio ? { aspectRatio: opts.aspectRatio } : {}),
+      ...(opts.seed !== undefined ? { seed: opts.seed } : {}),
+      providerOptions,
+    })
+  } catch (err) {
+    return toError(err)
+  }
+
+  const images = result.images.map((img) => ({
+    uint8Array: img.uint8Array,
+    mediaType: img.mediaType,
+  }))
+
+  const cost = calculateCost(config.cost, result.usage, result.images.length)
+
+  return {
+    images,
+    model: opts.model,
+    cost,
+    usage: result.usage,
+    warnings: result.warnings,
+  }
+}
+
+async function generateWithTextModelPath(opts: {
+  prompt: string
+  model: string
+  inputImages: Uint8Array[]
+  imageSize?: '1K' | '2K' | '4K'
+  aspectRatio?: string
+}): Promise<Error | GenerateImageResult> {
+  const textModel = await createTextModel(opts.model)
+  if (textModel instanceof Error) return textModel
+
+  const config = getModelConfig(opts.model)
+  if (config instanceof Error) return config
+
+  const messages = opts.inputImages.length > 0
+    ? [
+        {
+          role: 'user' as const,
+          content: [
+            { type: 'text' as const, text: opts.prompt },
+            ...opts.inputImages.map((img) => ({
+              type: 'image' as const,
+              image: img,
+            })),
+          ],
+        },
+      ]
+    : undefined
+
+  type GoogleAspectRatio = NonNullable<NonNullable<GoogleLanguageModelOptions['imageConfig']>['aspectRatio']>
+  const googleOpts = {
+    responseModalities: ['TEXT', 'IMAGE'],
+    ...(opts.imageSize || opts.aspectRatio
+      ? {
+          imageConfig: {
+            ...(opts.imageSize ? { imageSize: opts.imageSize } : {}),
+            ...(opts.aspectRatio ? { aspectRatio: opts.aspectRatio as GoogleAspectRatio } : {}),
+          },
+        }
+      : {}),
+  } satisfies GoogleLanguageModelOptions
+
+  const providerOptionsKey = config.provider === 'vertex' ? 'vertex' : 'google'
+
+  let result
+  try {
+    result = await generateText({
+      model: textModel,
+      ...(messages ? { messages } : { prompt: opts.prompt }),
+      providerOptions: {
+        [providerOptionsKey]: googleOpts,
+      },
+    })
+  } catch (err) {
+    return toError(err)
+  }
+
+  const imageFiles = result.files.filter((f) => f.mediaType.startsWith('image/'))
+  if (imageFiles.length === 0) {
+    return new Error('No images generated by the text model.')
+  }
+
+  const images = imageFiles.map((f) => ({
+    uint8Array: f.uint8Array,
+    mediaType: f.mediaType,
+  }))
+
+  const cost = calculateCost(config.cost, result.usage, imageFiles.length)
+
+  return {
+    images,
+    model: opts.model,
+    cost,
+    usage: result.usage,
+    text: result.text || undefined,
+  }
+}
+
+async function generateWithResponsesApi(opts: {
+  prompt: string
+  model: string
+  count: number
+  aspectRatio?: `${number}:${number}`
+  seed?: number
+  maskImage?: Uint8Array
+  inputImages: Uint8Array[]
+}): Promise<Error | GenerateImageResult> {
+  if (opts.count !== 1) {
+    return new ValidationError('ChatGPT image generation currently supports exactly one output image.')
+  }
+
+  if (opts.seed !== undefined) {
+    return new ValidationError('ChatGPT image generation does not support seed.')
+  }
+
+  if (opts.maskImage) {
+    return new ValidationError('ChatGPT image generation does not support mask yet.')
+  }
+
+  const size = opts.aspectRatio ? chatGptImageSizeFromAspectRatio(opts.aspectRatio) : undefined
+  if (opts.aspectRatio && !size) {
+    return new ValidationError('ChatGPT image generation only supports aspect ratios 1:1, 3:2, or 2:3.')
+  }
+
+  const storedAuth = getChatGptAuth()
+  if (!storedAuth?.accountId) {
+    return new ValidationError('Missing ChatGPT account metadata. Please run `egaki login --provider chatgpt` again.')
+  }
+
+  const auth = await getValidChatGptAuth(storedAuth, saveChatGptAuth)
+  if (auth instanceof Error) return auth
+
+  const content: Array<{ type: 'input_text'; text: string } | { type: 'input_image'; image_url: string }> = [
+    { type: 'input_text', text: opts.prompt },
+    ...opts.inputImages.map((image) => ({
+      type: 'input_image' as const,
+      image_url: imageBytesToDataUrl(image),
+    })),
+  ]
+
+  let response: Response
+  try {
+    response = await fetch('https://chatgpt.com/backend-api/codex/responses', {
+    method: 'POST',
+    headers: {
+      Authorization: `Bearer ${auth.access}`,
+      ...(auth.accountId && { 'ChatGPT-Account-ID': auth.accountId }),
+      'Content-Type': 'application/json',
+    },
+    body: JSON.stringify({
+      model: 'gpt-5.4',
+      instructions: 'You are Codex.',
+      input: [
+        {
+          type: 'message',
+          role: 'user',
+          content,
+        },
+      ],
+      tools: [
+        {
+          type: 'image_generation',
+          model: opts.model,
+          size: 'auto',
+          quality: 'auto',
+          output_format: 'png',
+          output_compression: 100,
+          moderation: 'auto',
+          ...(size ? { size } : {}),
+        },
+      ],
+      tool_choice: 'auto',
+      parallel_tool_calls: true,
+      stream: true,
+      store: false,
+      include: [],
+    }),
+  })
+  } catch (err) {
+    return toError(err)
+  }
+
+  if (!response.ok || !response.body) {
+    const body = await response.text().catch(() => '')
+    return new Error(`ChatGPT image generation failed: ${response.status}${body ? ` ${body}` : ''}`)
+  }
+
+  let imageBase64: string | undefined
+  let revisedPrompt: string | null = null
+
+  const parser = createParser({
+    onEvent: (message: EventSourceMessage) => {
+      if (!message.data) return
+      try {
+        const event = JSON.parse(message.data) as {
+          type?: string
+          partial_image_b64?: string
+          item?: { type?: string; result?: string; revised_prompt?: string | null }
+        }
+        if (
+          event.type === 'response.image_generation_call.partial_image' &&
+          event.partial_image_b64 &&
+          !imageBase64
+        ) {
+          imageBase64 = event.partial_image_b64
+        }
+        if (
+          event.type === 'response.output_item.done' &&
+          event.item?.type === 'image_generation_call'
+        ) {
+          if (event.item.result) imageBase64 = event.item.result
+          revisedPrompt = event.item.revised_prompt ?? revisedPrompt
+        }
+      } catch {
+        // Ignore keepalive and partial parse noise.
+      }
+    },
+  })
+
+  const decoder = new TextDecoder()
+  for await (const chunk of response.body as unknown as AsyncIterable<Uint8Array>) {
+    parser.feed(decoder.decode(chunk, { stream: true }))
+  }
+  parser.feed(decoder.decode())
+
+  if (!imageBase64) {
+    return new Error('No images generated by ChatGPT.')
+  }
+
+  const imageBytes = new Uint8Array(Buffer.from(imageBase64, 'base64'))
+
+  return {
+    images: [{ uint8Array: imageBytes, mediaType: 'image/png' }],
+    model: opts.model,
+    cost: null,
+    revisedPrompt,
+  }
+}
+
+// ─── cached result types ─────────────────────────────────────────────────────
+
+export interface CachedGenerateResult {
+  /** Public URL path (e.g. `/generated/image/sunset-a1b2c3d4.png`) */
+  src: string
+}
+
+// ─── cached generateImage ───────────────────────────────────────────────────
+
+import { cachedGenerate } from './cached-generate.js'
+import { extensionFromMediaType } from './cache-utils.js'
+
+export const generateImage = cachedGenerate<GenerateImageOptions & { seed?: number }, GeneratedFile, CachedGenerateResult>({
+  namespace: 'image',
+  prefixFrom: (p) => p.prompt,
+  modelFrom: (p) => p.model,
+  generate: async (params) => {
+    const result = await generateImageUncached(params)
+    if (result instanceof Error) throw result
+    const file = result.images[0]
+    if (!file) throw new Error('No image generated')
+    return file
+  },
+  serialize: (file) => ({
+    bytes: file.uint8Array,
+    extension: extensionFromMediaType(file.mediaType),
+  }),
+})
+
+// ─── cached generateVideo ───────────────────────────────────────────────────
+
+export const generateVideo = cachedGenerate<GenerateVideoOptions & { seed?: number }, GeneratedFile, CachedGenerateResult>({
+  namespace: 'video',
+  prefixFrom: (p) => p.prompt,
+  modelFrom: (p) => p.model,
+  generate: async (params) => {
+    const result = await generateVideoUncached(params)
+    if (result instanceof Error) throw result
+    const file = result.videos[0]
+    if (!file) throw new Error('No video generated')
+    return file
+  },
+  serialize: (file) => ({
+    bytes: file.uint8Array,
+    extension: extensionFromMediaType(file.mediaType),
+  }),
+})
+
+// ─── re-export speech generation ─────────────────────────────────────────────
+
+export {
+  generateSpeech,
+  generateSpeechUncached,
+  calculateSpeechCost,
+  type GenerateSpeechOptions,
+  type GenerateSpeechResult,
+  type CachedSpeechResult,
+  type SpeechProvider,
+  type SpeechProviderOptions,
+  type SpeechProviderResult,
+} from './speech-generate.js'
+
+// ─── re-export transcription ─────────────────────────────────────────────────
+
+export {
+  transcribeAudio,
+  transcribeAudioUncached,
+  calculateTranscriptionCost,
+  segmentsToWordTimestamps,
+  wordTimestampsToCaptions,
+  type TranscribeOptions,
+  type TranscribeResult,
+  type WordTimestamp,
+  type TranscriptionSegment,
+} from './transcription-generate.js'