runware

What This Skill Does

The Runware skill for OpenClaw integrates powerful generative AI capabilities directly into your agent workflow. By leveraging the Runware unified API, it provides programmatic access to top-tier models including FLUX, Stable Diffusion XL, and Kling AI. This skill enables the generation of high-quality static images from text, sophisticated image-to-image transformations, intelligent upscaling, and advanced video synthesis. Whether you are creating social media content, prototyping visual concepts, or animating static assets, the Runware skill offers a robust, model-agnostic interface that handles the heavy lifting of model selection, parameter configuration, and async polling.

Installation

To begin using this skill, ensure you have the OpenClaw environment properly configured. Install the package using the following command in your terminal: clawhub install openclaw/skills/skills/26medias/runware. Before executing any generation scripts, you must obtain a valid API key from https://runware.ai. Once you have your key, set it as an environment variable named RUNWARE_API_KEY for seamless authentication. Alternatively, you can pass the --api-key argument directly when executing the Python scripts located in the scripts/ directory.

Use Cases

This skill is designed for creative automation and media production. Use it to:

• Automate the generation of marketing assets using specific style LoRAs.
• Scale production of visual media by programmatically batch-generating images.
• Transform draft sketches or lower-resolution images into professional-grade visuals through img2img and upscaling.
• Animate static images or create cinematic clips from textual descriptions for video projects.
• Quickly iterate on AI model selection to find the best fit for your project's aesthetic, such as switching between FLUX.1 Dev for quality or Schnell for rapid prototyping.

Example Prompts

• "Generate 3 high-quality images of a futuristic cityscape at night in a synthwave art style using the FLUX.1 model."
• "Upscale the image located at ./input_draft.png by 4x and save the result as high_res_final.png."
• "Create a 5-second video of a sunset over a calm ocean based on this image: ./base_ocean.jpg, using the Kling AI model."

Tips & Limitations

When using video generation, remember that these tasks are asynchronous. The scripts will automatically poll the API until the rendering is complete, which may take several minutes depending on the complexity of the prompt and the load on the selected model. Keep in mind that costs are associated with each generation request, so monitor your usage on the Runware dashboard. For images, start with default parameters (25 steps, CFG 7.5) and adjust only if the output quality or composition doesn't meet your requirements. Always provide descriptive prompts with motion keywords for video generation to achieve the most accurate movement results. Note that the script requires file system access to read source images and write final outputs.