Image Resizer for AI Vision Inputs
Resize or center-crop images to the dimensions vision models actually use: 1568px long edge for Claude, 1024, 512, or any custom size. Preview the result and download as PNG or JPEG.
100% private: resizing happens in your browser with FileReader and canvas. Your image is never uploaded anywhere.
Drop an image here or click to pick one
Processed locally in your browser - nothing is uploaded
Processed locally in your browser - nothing is uploaded
Why size matters for vision APIs: vision models downscale big images anyway, but you still pay to transfer and, on some providers, to tokenize the larger input. Claude works best when the long edge is at most 1568 px; anything bigger is resized server-side, so the extra pixels are pure waste. A 4000 x 3000 photo sent raw can cost noticeably more per request and adds upload latency, while a properly sized 1568 px version reads exactly the same to the model. Rule of thumb: 1568 px for detailed screenshots and documents, 1024 px for general photos, 512 px for thumbnails or quick classification. Sizing guidance is approximate, as of July 2026, check provider docs. Pair this with the Image Compressor to cut file size further, and the Vision API Cost Calculator to see the dollar impact.