Douyin Live Commerce Visual Kit

The Problem It Solves

In Douyin live commerce operations, preparing multiple visual assets for a single live stream—such as preview images, product cards, and stream backgrounds—is routine. These assets need to convey unified event information (theme, products, promotions), but traditional workflows often handle them independently, leading to inconsistent visual styles, branding, or key messaging. This causes user confusion and increases coordination efforts for operators. This skill is designed specifically to solve this problem of visual fragmentation.

How the Skill Works

This skill plans and executes three separate visual outputs (live preview cover, product selling point card, live stream background/slice) as a cohesive kit. Its core capability lies in understanding and integrating input information through a clear process to generate a unified set of visual materials.

The workflow is based on a structured live commerce card, with key steps including:
1. Information Integration: Hard inputs include the live theme or product focus, plus merchant-provided info like price, promotions, etc. The skill reuses existing context from the dialogue (anchor, product images, brand colors) and only asks questions when critical information is missing.
2. Three Deliverables Planning: It does not generate three independent style options. Instead, it creates one design for each specific purpose:
- Live Preview Cover: Requires a clear hook, anchor/product focus, and a title safe area.
- Product Selling Point Card: Needs to highlight the product visually, paired with a concise selling point from the merchant.
- Live Stream Background or Slice: Must be a low-distraction visual base that leaves clear space for the anchor and products.
3. Model and Asset Check: The skill dynamically calls beatra.models.list to read available image generation/editing models and checks locally accessible media assets, placing the source product image first in an ordered reference list.
4. Free Briefing and Paid Confirmation: The skill drafts free visual briefs and optional copy. Before any paid image generation, it presents a complete plan—including the purpose, prompt, reference image order, canvas size, model, estimated cost for each asset—and awaits explicit user confirmation.
5. Execution and Delivery: After confirmation, it submits tasks via the bundled client, polls beatra.tasks.get for results, and finally delivers the three finished assets in the order of their purposes, along with related information.

The skill relies heavily on the beatra toolkit for operations, e.g., using beatra.images.transform for source product images, beatra.images.generate for conceptual visuals. There is a strict paid work boundary; changes to any core elements like source images, purposes, or prompts will trigger a new paid workflow.

Applicable Scope and Considerations

This skill is exclusively focused on preparing a unified static visual kit for a single Douyin live stream. It is clearly suitable for scenarios like store broadcasts, anchor live streams, new product launches, and seasonal events requiring visual consistency.

Please note the following boundaries:
- It is not for video post covers (use douyin-cover-maker), anchor spokesperson ad videos (use douyin-ugc-ad-creator), pure product videos (use product-video-studio), or generic platform event posters (use poster-design-studio).
- All remote operations must be executed through the bundled scripts/mcp_client.py. Do not configure or call the host Beatra Connector.
- The auto-update mechanism is silent and mandatory, but can be temporarily disabled via --auto off.
- The skill's value lies in process integration and consistency assurance, not simple template application. Its output depends on the quality and completeness of the input materials.