Drop images here
or click to browse — PNG, JPG, WEBP
Original
—
Optimised
—
Reduction
—
Downscale and convert to greyscale — reduces file size and lowers token costs when sharing images with AI tools. Runs entirely in your browser. Nothing is uploaded.
or click to browse — PNG, JPG, WEBP
Frequently asked questions
Yes. AI vision models process colour information across RGB channels. Converting to greyscale collapses that to a single luminance value per pixel, meaningfully reducing the encoded payload the model processes. Combined with downscaling, typical file size reductions of 60–80% are achievable, which directly lowers token cost.
No. This tool runs entirely in your browser using the HTML5 Canvas API. No image data leaves your device at any point — it is fully private and offline-capable.
For most reference and working images — screenshots, diagrams, mockups — 800px wide is a good default. The AI can still read text and identify content at this resolution whilst the token cost is significantly lower than a full-resolution image. For fine-detail analysis, try 1200–1600px.
Input: PNG, JPG and WEBP. Output: JPEG (with your chosen quality setting). Batch processing of multiple files is supported.