Hi Product Hunt! I'm Chen, the maker of Scenepond.

When you're making a sequence, each shot has context: the script it belongs to, the reference you're using, and the changes you still want to make. Scenepond is built around keeping that context together while you create.

It combines a conversational AI agent with a visual canvas. You can describe what you're trying to make, work through a shot plan, and keep developing the project on the canvas.

Here are three things you can do:

- Develop an idea or script into a storyboard and shot plan.

- Keep references, prompts, and generated images or clips together.

- Review and revise individual shots as the sequence develops.

The gallery follows Sky Whale, a real Scenepond project with character and scene references, shot-by-shot directions, and two 15-second video clips. The demo shows the actual canvas, saved takes, and a full generated clip.

For a first try, start with one reference image and one shot. Describe the motion, review the model settings and credit quote, then generate and inspect the result before expanding the project.

Scenepond supports English and Simplified Chinese. It's a paid product with trial credits after email verification; generation costs depend on the model and settings.

I'd love feedback from people making short videos and visual stories: when a shot needs another revision, what do you find yourself going back to look for?