Text-to-Video and Image-to-Video Studio

Specific Problem It Solves

In short video production, creators need to convert text descriptions, images, or existing videos into dynamic content quickly for ads, product stories, or social media. Traditional methods may involve multiple tools and complex workflows, making efficient iteration and effect verification difficult. This skill addresses this by providing an integrated workspace to generate short videos directly from materials, streamlining the path from idea to video.

How the Skill Works

The skill achieves video production through core capabilities and key steps:

• Six Video Routes: Select the optimal path based on input material type:
  1. beatra.videos.animate: Generate video using an image as the opening frame.
  2. beatra.videos.interpolate: Generate video by interpolating between start and end frames.
  3. beatra.videos.generate_from_references: Generate a new video based on ordered reference materials.
  4. beatra.videos.edit: Modify existing video content.
  5. beatra.videos.extend: Append frames before or after a video.
  6. Text descriptions can first generate a key frame with beatra.images.generate, then convert to video.

• Key Steps: First, check source material attributes like MIME type, dimensions, and duration. Then choose a video route and call beatra.models.list to get available model information. Next, present a video entry card, confirm payment, and submit the request using a stable client_request_id. Finally, poll task status (queued or running indicates in progress) and deliver results.

• Defaults and Optimization: Prioritize user-provided materials; if no key frame, generate an independent image stage first. The model defaults to model: "auto", with duration and resolution automatically selected based on model card for the shortest integer unless specified higher.

Applicable Boundaries and Considerations

• Non-Applicable Scenarios: This skill does not handle subtitle, voiceover synthesis, or timeline assembly; multi-shot videos must be delivered as independent videos. If only talking avatar or image animation is needed, use specialized talking-avatar or image-to-motion tools.

• Considerations: Each payment phase requires separate confirmation; changing parameters (e.g., model, duration) requires re-planning and presenting a new entry card. Task polling does not auto-retry, and after delivery, check if the video matches requested actions, camera movement, etc., then recommend the minimal effective next step (e.g., editing or extension). Do not imply an unchecked video is a finished promotional piece.