project-docs_skill

Overview

This skill documents the AI Content Pipeline for the video-ai-studio Python package and serves as an architecture and usage reference. It centralizes model registry details, CLI commands, YAML pipeline syntax, provider integrations, error codes, and implementation locations. Use it to quickly locate files, understand design decisions, and get practical examples for pipelines and integrations.

How this skill works

The skill maps documentation topics to source files and code locations so you can jump from a question to the exact doc or module (e.g., registry.py, cli/click_app.py, manager.py). It summarizes key facts: available models and categories, provider support, CLI commands, YAML pipeline features, and testing commands. It highlights where to find implementation details (package structure, architecture diagrams) and troubleshooting/error-code guidance. It also points to guides for optimization, cost management, and best practices for pipeline design.

When to use it

• Exploring the codebase structure and import paths for development or auditing.
• Choosing models or providers and comparing pricing and capabilities.
• Writing or debugging YAML pipelines, including parallel execution patterns.
• Learning CLI usage and available commands for generation, analysis, and project workflows.
• Integrating the pipeline into an application (Flask, FastAPI, Celery) or automating workflows.

Best practices

• Use the central registry (registry.py / registry_data.py) to find canonical model keys and metadata before hardcoding names.
• Prefer YAML pipelines for repeatable workflows; leverage parallel groups for throughput-sensitive steps.
• Store API keys in environment variables (.env) and follow the security guide for production deployments.
• Run tests and registry validation scripts during CI to catch breaking provider or model changes early.
• Use cost-management and performance guides to estimate expenses and tune batching/caching.

Example use cases

• Create a multi-stage pipeline that generates prompts, renders frames with a text-to-video model, then adds TTS and upscales output.
• Audit provider choices by comparing FAL AI, Google Gemini, ElevenLabs, OpenRouter, and Replicate for cost and latency.
• Automate media ops via CLI: transcribe, analyze-video, transfer-motion, and generate-avatar in a CI step.
• Prototype novel-to-video workflows using the ViMax subgroup commands (idea2video, novel2movie, script2video).
• Instrument a production executor with parallel execution to achieve 2-3x speedup on batch jobs.