ai-image-generator

What This Skill Does

The ai-image-generator is a powerful, asynchronous skill for OpenClaw that serves as a gateway to advanced AI media generation. It leverages the AI Artist API to convert textual prompts into high-quality images and video content. The skill intelligently manages the asynchronous nature of generative AI by implementing an automated polling mechanism that tracks task status until completion, ensuring users receive their assets seamlessly. It supports a diverse range of models, including SEEDREAM5_0 for high-fidelity images, NANO_BANANA_2 for efficient, lightweight generation, and specialized video models like SEEDANCE_1_5_PRO and SORA2, which support complex features like audio integration and first-frame/last-frame image-to-video guidance.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/2393970875/deepsop-image-video-generator.

Important Setup Steps:

1. Obtain an API Key from https://staging.kocgo.vip/index.
2. Set your API Key in the environment variable AI_ARTIST_TOKEN to ensure authenticated access to the generative services.
3. Verify your configuration by running the built-in diagnostic tool python3 scripts/test_config.py.

Use Cases

• Creative Visualization: Generate artistic interpretations of descriptive prompts for blog posts, social media, or presentations.
• Content Creation: Produce short-form video snippets (up to 10 seconds) with audio for marketing or entertainment purposes.
• Rapid Prototyping: Quickly iterate on visual concepts using the NANO_BANANA_2 model to test prompt effectiveness before rendering high-quality 4K assets.
• Media Integration: Automatically download or stream media generated by the agent into your workflow, supporting workflows that require local storage or direct Markdown-based rendering.

Example Prompts

• "帮我生成一匹在雪地里奔跑的狼，质量要高，使用 SEEDREAM5_0 模型。"
• "用 SORA2 模型帮我生成一段 16:9 的视频，内容是一只猫在窗台晒太阳，时长 10 秒。"
• "画一张 4K 分辨率的未来科幻城市风景图，要求细节丰富，使用最高质量配置。"

Tips & Limitations

• Model Selection: Choose your model based on your needs. Use NANO_BANANA_2 for quick previews, but switch to SEEDREAM5_0 for production-grade, high-resolution imagery.
• Video Capabilities: Note that SORA2 and SEEDANCE_1_5_PRO have different capabilities; ensure you provide the correct parameters such as firstImageUrl or lastImageUrl when leveraging advanced video features.
• Data Storage: By default, files are saved in workspace/images. Ensure your system has sufficient disk space if you frequently use the --download flag. Always verify if your current platform supports direct Markdown image display, as some messaging apps (like WhatsApp) have limitations.