> ## Documentation Index > Fetch the complete documentation index at: https://docs.akool.com/llms.txt > Use this file to discover all available pages before exploring further. # Create Video Translation > Translate videos into multiple languages with AI voices and lip-sync Create a video translation job to translate your video content into one or multiple target languages. The API supports automatic source language detection, automatic voice selection, lip-sync, and various caption options. > [!NOTE] > > The `merge_interval` and `face_enhance` parameters have been deprecated and are no longer supported. ## Voice Selection The API selects the voice type separately for each target language: * **Original voice clone (`voice_type: 0`)**: Used when the target language exactly matches one of the following codes: `en`, `fr`, `zh`, `ar`, `bg`, `hr`, `cs`, `da`, `nl`, `fil`, `fi`, `de`, `el`, `hi`, `hu`, `id`, `it`, `ja`, `ko`, `ms`, `nb`, `pl`, `pt`, `ro`, `ru`, `sk`, `es`, `sv`, `ta`, `tr`, `uk`, `vi`. * **AI voice (`voice_type: 2`)**: Used for target language codes outside this list. Matching uses the exact language code. For example, `zh` selects original voice cloning, while `zh-CN` selects AI voice. For comma-separated targets, use language codes without surrounding spaces. You can provide optional voice settings in `voices_map`, keyed by target language. The server overrides any supplied `voice_type` according to the rules above and passes the resulting settings to the translation service. This endpoint does not read an `ai_voice` request parameter. A `voice_id` is not required by the current OpenAPI request validation. ## Video Duration `/api/open/v3/content/video/createbytranslate` does not enforce a 60-second duration limit in the current OpenAPI validation. Shorter videos may reduce processing time; keeping videos under 60 seconds is a recommendation, not a request requirement. Downstream processing constraints may still apply. The separate legacy endpoint `/api/open/v3/content/video/translate` rejects videos longer than 60 seconds. This restriction does not apply to the Create Video Translation endpoint documented here. ## OpenAPI ````yaml POST /api/open/v3/content/video/createbytranslate openapi: 3.0.3 info: title: Video Translation API description: >- API for translating videos into multiple languages with AI voices and lip-sync version: 1.0.0 servers: - url: https://openapi.akool.com description: Production server security: - ApiKeyAuth: [] - BearerAuth: [] paths: /api/open/v3/content/video/createbytranslate: post: tags: - Video Translation summary: Create Video Translation description: > Translate videos into multiple languages with automatic voice selection and lip-sync. The server sets voice_type to 0 for target language codes in the original voice clone list and to 2 for other target languages, overriding any supplied voice_type. This endpoint does not read an ai_voice request parameter. Current OpenAPI validation does not enforce a 60-second duration limit. Shorter videos are recommended for faster processing; downstream processing constraints may still apply. The separate legacy /api/open/v3/content/video/translate endpoint rejects videos longer than 60 seconds. requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/VideoTranslationRequest' example: url: >- https://d11fbe263bhqij.cloudfront.net/agicontent/video/translate/cut3_content_create_EN_01.mp4 language: ja,zh-CN-shandong,zh-CN-sichuan source_language: en lipsync: false speaker_num: 1 webhookUrl: '' remove_bgm: true caption_type: 2 caption_url: >- https://drz0f01yeq1cx.cloudfront.net/1759979948074-b14aeb874a4547ca81373d140edd02d2-1758273867083fa8fa1ca7cfe409ab7fb07abdb931660translatedsubtitles.srt studio_voice: no_translate_words: - kaka - lal fixed_words: hi: hello bad: good pronounced_words: lel: lele gga: gaga style: professional dynamic_duration: false responses: '200': description: Video translation request submitted successfully content: application/json: schema: allOf: - $ref: '#/components/schemas/ApiResponse' - type: object properties: data: $ref: '#/components/schemas/VideoTranslationResult' components: schemas: VideoTranslationRequest: type: object required: - url - language properties: url: type: string description: >- The video URL address you want to translate. Current OpenAPI validation does not enforce a 60-second duration limit; shorter videos are recommended for faster processing. source_language: type: string default: DEFAULT description: >- The original language of the video. Use "DEFAULT" for automatic language detection, or specify a language code (e.g., "en" for English, "zh" for Chinese, "es" for Spanish) language: type: string description: > Target language(s) for translation. For multiple languages, separate with commas without surrounding spaces (e.g., "ja,zh-CN-shandong,zh-CN-sichuan"). Exact matches to en, fr, zh, ar, bg, hr, cs, da, nl, fil, fi, de, el, hi, hu, id, it, ja, ko, ms, nb, pl, pt, ro, ru, sk, es, sv, ta, tr, uk, vi use original voice cloning (voice_type 0). Other target codes use AI voice (voice_type 2). For example, zh uses voice_type 0, while zh-CN uses voice_type 2. lipsync: type: boolean default: false description: >- Enable synchronized mouth movements with the audio track in translated video webhookUrl: type: string description: >- Callback URL address for HTTP notifications when processing completes speaker_num: type: integer default: 0 description: Number of speakers in the video (0 = Auto Detect) remove_bgm: type: boolean default: false description: Whether to remove background music from the video caption_type: type: integer default: 0 enum: - 0 - 1 - 2 - 3 - 4 description: >- Caption handling: 0=none, 1=add original subtitle, 2=add target subtitle, 3=translate and replace original subtitle, 4=add translated subtitle caption_url: type: string description: Subtitle file URL (supports SRT or ASS files) voices_map: type: object description: >- Optional voice settings keyed by exact target language code. The server sets the language and overrides voice_type for each target according to the language selection rules. The current OpenAPI validation does not require voice_id. additionalProperties: type: object properties: voice_id: type: string description: Optional voice identifier passed to the translation service. voice_type: type: integer description: >- Overridden by the server with 0 (original voice clone) or 2 (AI voice), based on the target language code. studio_voice: type: object description: Advanced voice and translation settings properties: no_translate_words: type: array items: type: string description: List of words that should not be translated style: type: string default: professional enum: - affectionate - angry - calm - cheerful - depressed - disgruntled - embarrassed - empathetic - excited - fearful - friendly - unfriendly - sad - serious - relaxed - professional description: Translation style fixed_words: type: object additionalProperties: type: string description: Translation word mapping table for corrections pronounced_words: type: object additionalProperties: type: string description: Pronunciation correction word mapping table dynamic_duration: type: boolean default: false description: Control the dynamic duration of the video ApiResponse: type: object properties: code: type: integer description: Business status code (1000 = success) msg: type: string description: Status message VideoTranslationResult: type: object properties: _id: type: string description: Unique task identifier video_status: type: integer description: >- Video processing status: 1=queueing, 2=processing, 3=completed, 4=failed video: type: string description: Generated video URL (available when status = 3) securitySchemes: ApiKeyAuth: type: apiKey in: header name: x-api-key BearerAuth: type: http scheme: bearer ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.