chanjing-text-to-digital-person

What This Skill Does

The chanjing-text-to-digital-person skill empowers users to generate AI-driven digital characters directly through the Chanjing API suite. It acts as an orchestrator for a multi-stage creative process: transforming text descriptions into lifelike portraits, animating those portraits into talking videos, and managing specialized LoRA model training. Designed for developers and content creators, it simplifies complex asynchronous workflows by handling task polling, status verification, and final asset management, ensuring seamless integration into your media generation pipeline.

Installation

To install this skill, use the following command in your terminal: clawhub install openclaw/skills/skills/binkes/chanjing-text-to-digital-person Ensure you have configured your credentials via the chanjing-credentials-guard before execution. If not authorized, the skill will prompt you to open the login portal to authenticate your API keys.

Use Cases

• Marketing Automation: Rapidly generate diverse presenter characters for educational content or brand promotions based on specific demographic and industry requirements.
• Content Localization: Create consistent digital avatars that can be repurposed for talking-head videos with varying scripts.
• Personalized AI Training: Use the LoRA capabilities to fine-tune specific character styles or likenesses for highly targeted branding requirements.
• Asset Batching: Automate the entire lifecycle from creative ideation (text) to production (video) for large-scale media projects.

Example Prompts

1. "Generate a realistic professional woman for an educational video. Use a modern office background and make her look like she's standing and explaining concepts."
2. "Take the last digital character I generated and create a talking video where she speaks with a confident, clear tone. Download the result to my local folder."
3. "Train a new LoRA model using these five portrait photos provided in the attachments to capture a specific brand mascot style."

Tips & Limitations

• Workflow Logic: Always follow the standardized workflow (create task -> poll task -> download) to avoid handle failures due to the asynchronous nature of AIGC APIs.
• Task Distinction: Remember that list_tasks categorizes assets as type 1 (photos) or type 2 (videos). Use this to keep track of your generated media.
• LoRA Pre-processing: Ensure you provide high-quality, diverse image sets when training LoRA for better character fidelity. If you need to animate existing actors, consider using the chanjing-video-compose skill instead for specialized lip-sync tasks.