WeChat Channels Cover Maker Tool

Problem Context

When creating videos for WeChat Channels, designing an effective cover image often involves a delicate balance between highlighting the main subject and leaving adequate space for the title text. Whether for knowledge sharing, product explanations, or brand promotion, a clear cover directly impacts the first impression and click-through rate. However, designers or creators frequently need to rapidly generate a cover that aligns with the account's style and title safety zones, based on diverse inputs like topics, scripts, or keyframe screenshots. Manually adjusting canvas ratios, masking, or compositing is time-consuming and makes it difficult to ensure visual legibility at small sizes.

Core Capabilities and Workflow

This skill offers three automated paths tailored to different input scenarios:
- Generate from Content: When you only have a topic, title, or script and do not need to preserve source images, use beatra.images.generate. This text-driven path requires a clear visual direction and cannot be initiated with just a style word.
- Transform from Images: If you have exported keyframes, portraits, product photos, or reference images, use beatra.images.transform. This is the primary route for working with existing visual assets, such as converting a video screenshot into a formal cover. Note that it cannot read video files directly; you must provide a keyframe image as input.
- Refine Selected Cover: For localized edits on an existing cover, such as adjusting a region or adding elements, use beatra.images.edit. You are limited to modifying at most two normalized areas.

The workflow begins with compiling a cover brief. The system reuses all relevant context from your conversation, such as the title, account style, and target audience. The key input is the planned placement of the title text—typically prioritized in the upper or lower third of the canvas to leave space, rather than covering the entire image. Next comes the planning stage, where you define the core subject, a small-scale visual hook, and information hierarchy. Before executing paid operations like image generation, transformation, or editing, you must confirm and freeze all parameters, including the precise canvas dimensions, prompts, image roles and sequence, model selection, and billing details. Any parameter change will initiate a new paid workflow.

After generation, the skill polls the task status via beatra.tasks.get, allowing you to track progress. Delivery includes the image link, observed dimensions, task ID, and billing breakdown. The entire process strictly uses the bundled scripts/mcp_client.py and does not employ other interfaces.

Scope and Important Notes

Please note the following core boundaries of this skill:
- It focuses solely on producing static cover images. It does not automatically extract frames from videos or handle video publishing.
- It is designed for general content covers on WeChat Channels. For WeChat Official Account article covers, use wechat-cover-maker; for WeChat Channels product videos, use wechat-channels-product-video. To evaluate an existing cover's performance only, use cover-performance-preflight.
- The canvas ratio must be explicitly confirmed. Do not default to the common 3:4 ratio used on platforms like Xiaohongshu; decide based on the actual placement location or source image ratio. When using image transformation, the last image in the sequence anchors the canvas aspect ratio—you must specify this role clearly.
- All image generation, transformation, and refinement operations are paid. Before submitting a request, carefully review the frozen parameter list, which includes a unique client_request_id, to prevent duplicate billing.
- The skill automatically updates the client at runtime but ensures safety and rollback mechanisms, so update failures do not affect the current task execution.