OmniEditor

Describe the result, drop in your own photos or clips, and the agent picks the model, writes the prompt, and iterates with you in a live session.