ShimmerPic

Hi Fazier! I built ShimmerPic because I kept switching between tabs, API keys and subscriptions just to see which image model handled a brief best. Now it's one studio with 9 models: GPT Image 2.5 Flare and Sunburst, GPT Image 2, the Nano Banana family, Seedream 5.0 Pro, Wan 2.7 Image and Midjourney. The newest update lets you keep editing a result step by step: open any image as a reference, describe the change, branch from an older version, and compare two versions side by side. You always see the exact credit cost before you generate, and everything you make stays private. New accounts get 30 free credits to try every model. I'd love to hear which model works best for your kind of work, and what you'd want next. (Independent tool, not affiliated with OpenAI.)

Trying the same idea on several models before choosing one is the use case that stands out to me — most studios lock you into a single model. The exact credit quote before generating is also a good call, since per-model pricing is usually opaque. One question: when you branch an older version, does the branch keep the original model, or can you switch models mid-edit?

Having nine models in one studio is useful because the best model often depends on the product and visual style. Showing the credit cost before generation and keeping creations private are strong practical choices. When refining an image, can users preserve the same product details while switching models, or does each model start a new interpretation?

This must be a useful product for social media posting

Interesting positioning. Putting several strong image models into one studio solves a very practical problem: constantly switching tools, subscriptions and workflows just to compare outputs. The step-by-step refinement and side-by-side comparison are especially useful — it feels more like a working environment than just another image generato