I have been working on Soutine, a web-based creative workspace that brings several AI image and video models into one place.

Create AI images and videos with Soutine

The project started because I was tired of switching between different tools whenever I wanted to generate an image, edit a reference photo, or turn a still image into a short video. Each service had its own interface, settings, and history. I wanted a simpler workflow where the input, model selection, results, and previous creations stayed together.

Soutine currently supports:

• Generating images from text prompts
• Editing and transforming reference images
• Creating videos from text or still images
• Upscaling videos and restoring visual detail
• Comparing different AI models in the same workspace

The available model pages include options such as Nano Banana, Nano Banana Pro, and GPT Image 2.5. Each model is suited to a slightly different job. Some are faster and less expensive, while others are better for detailed output, typography, reference-image editing, or higher resolutions.

I also added prompt libraries for people who prefer starting from an example instead of writing everything from scratch. You can browse a prompt, copy it, adjust the details, and open it in the creation workspace.

There are still rough edges. AI generation can be unpredictable, and the same prompt does not always produce the same quality across models. Part of the work now is making those differences easier to understand without turning the interface into a wall of technical settings.

For crawlers, AI tools, and anyone who wants a quick overview of the public pages, I published a plain-text site index:

Read the Soutine site index and model directory

Soutine is an independent project, and I am continuing to improve the model selection, generation workflow, and result history. If you work with AI images or videos, I would appreciate feedback on what feels useful, confusing, or unnecessarily complicated.