alicloud-ai-image-qwen-image

What This Skill Does

The alicloud-ai-image-qwen-image skill provides a robust interface for integrating high-fidelity image generation into the OpenClaw video-agent pipeline. Leveraging Alibaba Cloud’s Model Studio DashScope SDK, this skill enables the generation of high-quality visuals using the advanced Qwen Image models, specifically qwen-image-max and qwen-image-plus-2026-01-09. It acts as a bridge, normalizing requests to ensure consistent behavior regardless of the underlying API complexities. By standardizing input parameters like prompts, negative prompts, dimensions, and seeds, developers can achieve reproducible and consistent creative outputs, essential for automated video production workflows.

Installation

To get started, ensure you have the OpenClaw environment initialized. Install the skill directly via the CLI: clawhub install openclaw/skills/skills/cinience/alicloud-ai-image-qwen-image.

Before execution, ensure your environment is prepared by installing the DashScope Python SDK. It is recommended to use a virtual environment to manage dependencies properly:

1. Create and activate your venv: python3 -m venv .venv && . .venv/bin/activate
2. Install dependencies: python -m pip install dashscope
3. Set your authentication key: Export DASHSCOPE_API_KEY in your terminal or configure it in ~/.alibabacloud/credentials under the [default] section for seamless integration.

Use Cases

This skill is ideal for AI-driven creative pipelines. Use it to generate consistent assets for video storyboards, creating custom thumbnails, or generating character references based on specific prompts. It is particularly effective when integrated into a larger agentic workflow where a text-to-video agent needs supplementary high-resolution keyframes or style-matched background art to maintain thematic consistency throughout a production.

Example Prompts

1. "Generate a cinematic 1024x1024 portrait of a cyberpunk detective standing under neon rain, sharp focus, highly detailed."
2. "Create a professional product image of a sustainable water bottle on a white background, size 768x1024, minimalist aesthetic."
3. "Generate an abstract digital oil painting of a forest at sunrise using the qwen-image-max model, size 1024x1024, focus on warm orange and purple hues."

Tips & Limitations

To maximize the quality of your outputs, always provide descriptive, scene-focused prompts rather than simple keywords. While the negative_prompt field is supported, please note that it acts as a best-effort constraint; depending on the model's training, some negative cues may be ignored. For reproducible results, always specify a seed value. If you are integrating this into a high-throughput pipeline, ensure your DashScope API rate limits are configured appropriately to avoid service interruptions. Lastly, always verify your image size format matches the WxH specification (e.g., 1024*1024) to prevent request errors.