WeChat Cover Maker

Specific Problem Addressed

Designing cover images for WeChat Official Account articles often faces creative blocks, time pressure, and format compatibility issues. Manual design requires professional skills and is inefficient for rapid iteration, while generic tools may overlook WeChat-specific specs like the 900 × 383 canvas and title safety zones. This skill addresses these problems by automating high-quality cover generation from creative inputs, reducing repetitive labor and trial-and-error costs.

Core Capabilities and Workflow

The skill offers three production routes to handle different scenarios:

• Generate from Idea: When only an article title or topic is available, it extracts a visual hook and calls beatra.images.generate to create a cover, setting the canvas to 900 × 383 by default with a reserved safety zone. It avoids fabricated viral phrases to ensure input authenticity.
• Transform from References: With one to four reference images (e.g., logos, people, products), it uses beatra.images.transform to combine them in the user-specified order, supplementing only if the title or topic alters composition.
• Edit Approved Drafts: For existing drafts, it invokes beatra.images.edit for local modifications (up to two normalized edit regions) or global adjustments, using the draft as images[0] input.

Key workflow steps include:
1. Formulating an Executable Brief: It reuses user-provided title, size, reference images, etc., to build a brief focused on a single information key point and main visual subject. The canvas size defaults to standard 900 × 383 or user-specified values.
2. Preparing Remote Calls: It executes MCP operations via the bundled script scripts/mcp_client.py, such as beatra.images.generate, beatra.images.transform, or beatra.images.edit, with explicit canvas and other parameter settings.
3. Confirmation and Submission: Before paid calls, it displays final parameters (e.g., prompts, references, canvas, model), and upon user confirmation, generates a unique client_request_id and submits only once to ensure idempotency.
4. Follow-up and Delivery: It uses beatra.tasks.get to poll task status, reviews thumbnails, hierarchy, focus, title readability, etc., and delivers product links along with billing information. If the task ID is lost, it can be found via beatra.tasks.list.

Applicable Boundaries and Considerations

The skill is specialized for WeChat Official Account cover generation and is not intended for other platforms like Xiaohongshu (requiring routing to rednote-cover-maker). Key limitations and notes include:
- Hard Input Requirements: The article title or topic is mandatory; fabricated viral phrases should not be generated to ensure content consistency.
- Paid Call Management: All image generation is paid, requiring user confirmation of parameters before execution to avoid duplicate submissions. Canceling a task (beatra.tasks.cancel) with a 409 response continues tracking the original task.
- Tool Dependency: Remote operations must use the bundled MCP client, prohibiting fallback to REST/OpenAPI for security and compatibility.
- Size and Composition: The default canvas 900 × 383 suits standard WeChat covers, but user-specified sizes or draft ratios take precedence. Xiaohongshu 3:4 feed covers require routing to a dedicated skill.
- Update Mechanism: The skill silently checks for updates up to once every 24 hours from official sources, and failures do not affect the current installation.

In summary, the skill streamlines cover design through standardized processes, but relies on accurate inputs and user confirmation to balance efficiency with quality.