WorkRally CLI Tool

Why a CLI for AIGC Workflow?

Creating comics or animated series with AIGC involves a complex pipeline: managing project hierarchies, uploading and organizing vast media assets, orchestrating multiple AI models for image, video, and audio generation, and assembling outputs on an infinite canvas. Manually calling underlying APIs or relying on a web UI for batch operations is inefficient, error-prone, and difficult to automate. The WorkRally CLI addresses this by encapsulating over 30 core platform capabilities into a coherent, composable command-line interface, specifically optimized for programmatic use by AI Agents.

Core Capabilities and Workflow

The CLI's strength lies in decomposing a creative workflow into deterministic commands and steps. The process is generally grounded in the principles of "assets first" and "generation later."

Asset management is the foundation. It strictly differentiates between two containers:
* Media Library (asset): A project-level pool for raw files. All local or external files must first be uploaded here via workrally upload to obtain an official WorkRally URL. This is the only legitimate source for assets referenced in subsequent operations like AI generation.
* Asset Library (material): Provides a logical, tree-structured directory (e.g., for characters, props, scenes). Materials here can only be "mounted" from the Media Library, not uploaded directly.

AI Generation and Shot Creation are the core creative steps. After creating a shot, you define generation content by writing image_prompt, animation_prompt, and audio_prompt. Critical steps include:
1. Dynamic Model ID Fetching: Use commands like image-models and video-models to query available models in real-time. Hardcoding model IDs is strictly forbidden.
2. Pre-generation Config Check: Before issuing a generation command, always use shotlist get to verify that the target model is correctly set in the shot's extra.gen_config.
3. Execution and Status Polling: Generation commands only submit a task, not return results. You must use shotlist get-result to poll for and retrieve the final canvas node information.

Scope and Key Caveats

The WorkRally CLI is a deep integration tool for AI Agents, not a desktop app for end-user creators. Its effective use relies on understanding:

• Strong Dependency on API Concepts and Asset Logic: The tool is tightly coupled with WorkRally's concepts (Project, Canvas, Media Library, Asset Library). The most common failures stem from not following the "Upload -> Media Library -> Mount" asset pipeline or confusing the IDs of canvas and project.
• Designed for Reproducibility and Automation: The default output is JSON (-o json) for easy parsing by Agents. Critical commands like build-draft support incremental merging and concurrent edits, ensuring reliability in automated workflows.
• Strict Adherence to URL and Naming Conventions: Any parameter requiring a URL (e.g., --*-assets for generation) must be a WorkRally official URL obtained from workrally upload or asset get/search. Furthermore, the --name for assets should follow conventions like "canvasName_feature" or "promptKeyword" to facilitate later retrieval.

In essence, the WorkRally CLI codifies the entire process of comics/animation series creation, serving as foundational infrastructure for building automated, programmable AIGC content pipelines.