Upscale API

Upscale API uses AI super-resolution models to enhance the resolution of images and videos. Send a media URL with a scale factor (2x or 4x), and the model reconstructs detail at the higher resolution rather than simply interpolating pixels. The result is a genuinely sharper image or video, not a blurry enlargement. Image upscaling typically processes in 2-5 seconds; video upscaling depends on duration. Developers integrate upscaling into media processing pipelines, print preparation workflows, and content delivery systems. A publishing platform can accept low-resolution user uploads and automatically upscale them for print. A video platform can offer resolution enhancement as a premium feature. Archive digitization projects use it to restore detail in scanned historical documents and photos. For images, the best results come from inputs that are at least 256x256 pixels -- extremely small thumbnails may not have enough information for the model to reconstruct. For video, 2x upscaling is recommended over 4x to balance quality and processing time. Use the cost estimate endpoint first for video upscaling, as costs scale with duration and resolution.