PromptSketch AI – Image to Prompt

Generate structured, editable AI image prompts from web images, screen captures, and local files

Turn a reference image into an editable prompt without leaving the page. PromptSketch AI helps designers and AI creators understand visual references and reuse their creative direction. Start with a web image, a selected page area, a pasted image, or a local file. Review the result beside the reference and edit it before use. CHOOSE A REFERENCE • Right-click an image and select PromptSketch AI • Use the compact Prompt control when available • Detect the main visible image or capture a specific page area • Paste, drag, or load common image files • Add up to 10 references to one queue REVIEW AND REUSE • Get structured details about the subject, composition, color, lighting, style, and visible elements • Choose output languages before generating • Review a recreation prompt, prompt core, negative prompt, style tags, and structured data • Edit, copy, download, or export the result • Keep recent results in local history • Export history as JSON or Excel with image previews beside the prompts PROCESSING OPTIONS On supported devices, on-device processing can keep image analysis inside Chrome. Cloud Assist uses visual measurements and limited page context extracted by the extension without uploading the original image. You can also connect a compatible vision endpoint; selected images and requests go directly to that endpoint. Custom API credentials are stored in Chrome extension local storage. Operational analytics do not include images, prompt text, page bodies, clipboard content, or API keys. No account is required. The free version includes 20 Cloud Assist generations per day. Supported on-device processing and your configured vision endpoint remain available separately. Support: hello@reforgeextensions.com