Fiddl.art

Fiddl.art is a platform that lets users create AI images and videos by describing ideas to an agent, who selects models, writes prompts, and oversees cost and generation. The system presents plans and costs before creation and supports iterative refinement with the user. It emphasizes private creation by default and the ability to publish to the community if desired.

Users describe what they want in natural language or provide references, and the agent helps narrow the direction, find suitable models, and write prompts. The process includes reviewing the plan and cost before generation, and the agent can adjust light, outfits, or mood based on feedback.

The tool supports various use cases such as turning photos into edited images, creating illustrated versions, and generating short videos from starting images. It provides examples and prompts to guide users through typical scenarios, including headshots, remakes, and decorative designs.

The platform describes how costs, model choices, and outputs are reviewed and approved by the user before creation. It also notes that agents write prompts and select models suitable for the task and that users can request specific models or switch to manual control if needed.

Use cases

• Turn a photo into an edited image or a short video from a starting frame
• Generate themed visual content for social, ads, or listings