Reverse Prompt

Reverse Prompt is an image editor prompt tool. Reverse Prompt analyses a reference video or image and returns the reusable prompt behind it, including subject, style, shot structure, key frames and negative prompts.

Reverse Prompt is a web tool built around one idea: instead of writing a video prompt from scratch, you feed the system a reference and it writes the prompt for you. The input can be a still image, a video file, or a direct media link, and the tool accepts JPG, PNG, WebP, MP4 and MOV formats along with URLs. From that single upload, the AI reads the subject, the visual style, the shot structure and the negative prompts, then returns them as a structured prompt you can reuse in your own generation workflow. This is aimed at people who already make AI video or image content and need to match an existing look rather than invent one. Common use cases shown on the site include product shots, character work, poster design and short video clips. A marketer who wants a product ad in the same visual language as a competitor's spot, or a creator trying to replicate a specific aesthetic across a series, can upload the reference and get the descriptive language back instead of guessing at wording. The output goes beyond a single block of text. The tool breaks the reference into key frames, describes the shot structure, and produces both a positive prompt and a list of negative prompts, which matters for models that support exclusion terms. There is also an auto detect mode that decides whether the file is a video or an image, and a sample option if you want to see the output format before uploading your own media. Pricing is separate from the free upload flow. The site advertises early launch pricing based on workflow credits that cover prompts, images and videos, so heavy users should check the pricing page before committing. Compared with generic prompt libraries, the value here is specificity: the prompt describes your exact reference rather than a generic style category. It is a research and extraction step, not a generator, so it complements rather than replaces a video model.