Maestro is a video production service centered on an AI director. After describing the topic, audience, and communication goals, it organizes scripts, visuals, voiceovers, music, subtitles, and rendering to turn ideas into complete videos. It is suitable for educational explainers, product promotion, and multilingual content creation, and can also incorporate image, video, and audio assets to continue editing or extending existing projects.
Clarify capacity, inputs and outputs, and calling methods before choosing a model.
The above are the creation and API specifications for the Maestro video production entry point; the target duration does not equal the exact final video duration.
Learn what maestro can bring to your work.
Maestro does more than generate visuals: it connects topics, scripts, narration, music, subtitles, and editing into a production workflow. Prompts can clearly specify who the audience is, what needs to be explained, how the opening should capture attention, and how the ending should conclude, keeping creation focused on communication goals and making it suitable for starting production directly from a content brief.
Specify languages through langs, with the first serving as the primary language. Other languages reuse the same visuals before corresponding voiceovers and rendering are completed. Chinese, English, Japanese, and other versions can be organized in the same task, reducing the work of repeatedly preparing visual content; delivery results are separated by language for easier individual publishing.
There is no need to start from scratch after a video is completed. remix preserves the theme while adjusting the presentation, edit handles local changes such as titles, voiceovers, or color schemes, and extend expands the content. Provide a historical task ID and clear modification requirements to start a new iteration task and progressively refine the same video project.
Start with specific tasks to find where the model can be effective.
Enter a concept, tutorial key points, or organized article content, specify the audience's background and the conclusions you want to retain, then choose narrated to create an explainer video. The deliverable includes visuals, narration, and subtitles, making it suitable for turning text content into easy-to-watch short videos; key terms and required information should be specified in the prompt.
Attach reference assets such as product images and logos, describe the selling points, audience, and desired visual character, then choose a landscape, portrait, or square aspect ratio. Use presets such as modern and luxury to shape the look and feel, and pair them with a narration voice to support the message; after the video is created, continue modifying the title or voice-over to produce different promotional expressions.
When adding subtitles to existing material, use captions and submit the source video URL; when cross-language distribution is needed, select the target language in langs. The former focuses on processing existing videos, while the latter focuses on multilingual delivery of the same visual content, and they can be used respectively for asset organization and multi-region release of product introductions.
Choose based on task complexity, input materials, and expected results.
If the task requires not only visuals but also a script, voice-over, music, and subtitles, Maestro's director-style workflow is better suited to the goal of complete delivery. Rather than making requests only around a single visual segment, prioritize describing the content structure and audience expectations here. Use generate for new projects; for projects already created, choose an action based on whether you want to reinterpret, make local edits, or continue writing.
Choose narrated for explainer content, captions for adding subtitles to existing videos, avatar for talking-head videos, and drama for character dialogue stories. The type determines the form of expression, style adjusts the visual look and feel, and voice adjusts the narration voice; do not treat all three as a single control. If you want the system to organize the format itself, you can keep auto and clearly state the creative goal.
From a small-scale task to formal integration.
Define the goal, required inputs, and output requirements, using real business examples as a starting point.
Open the trial page, confirm the parameters supported by this entry point, then submit a small-scale task to review the results.
Keep the complete model ID, use the request format specified in the documentation, and confirm billing rules on the Pricing page.
Before using it officially, understand the output quality and capability scope.
Answers to common questions about using maestro.
Maestro is designed to organize complete video production. In addition to visuals, it includes scripts, narration, music, captions, editing, and rendering. Prompts should describe the content goal and presentation structure, rather than only the appearance of shots; it is better suited for creative tasks that move from a brief to a finished video.
Put media URLs in file_urls and explain the purpose of each asset in the prompt, such as using product images for display, logos for brand identification, and videos for caption processing. You can submit up to 20 items; when selecting captions, provide a source video, and when selecting avatar, provide a portrait.
You can specify up to 4 languages per request, with Chinese as the default; the first language in langs is the primary language. The remaining languages reuse the same visuals, with separate voice-overs and rendering, and language versions are provided upon completion. If you want each region to use different visual storytelling, submit separate creative requirements for each.
If you want to keep the theme but change the presentation, choose remix; to change specific elements such as the title, voice-over, or color scheme, choose edit; to continue expanding existing content, choose extend. All three require ref_task_id. Describe the modification goal in the prompt, and a new task ID will then be returned.
After calling POST /maestro/videos, first obtain task_id, then use POST /maestro/tasks to check progress and results. The task goes through planning and production stages; once successful, you can obtain the finished video and each language version. Applications should distinguish between submission successful, in production, completed, and failed states.
Model information · Updated: 2026-10-01. For call parameters and billing rules, see the API and pricing sections.
Start with a clear goal and evaluate whether it fits your work based on real results.
