The deep-fried image tool applies repeated JPEG compression passes to a photo, layered with color-channel shifting, to recreate the blocky, oversaturated look that images pick up after being shared and re-saved across social media dozens of times. It processes up to 20 images at once, outputs JPEG files, and runs entirely in the browser without uploading anything.
How Frying Strength and Quality Interact
Two parameters drive the effect:
Frying strength (iterations): 1–50 passes of JPEG re-encoding. At 1–5 the image still looks mostly normal. Around 10–20 you get visible blocky artifacts. Beyond 30 the image becomes heavily distorted. More iterations mean longer processing time.
Compression quality: 1–100, where lower means more damage per pass. At the default of 50 with 10 iterations you get a noticeable but readable result. Setting quality to 10–20 and iterations to 30+ produces extreme destruction.
The two interact differently: quality 50 + 30 iterations produces an even, uniform block pattern, while quality 20 + 10 iterations creates sharper high-frequency noise. Both reach similar levels of visual damage through different paths.
Maximum fry
- Iterations: 30-50
- Quality: 10-25
- Color: Green (classic deep-fried look)
Light vintage effect
- Iterations: 5-8
- Quality: 60-70
- Color: None or Yellow
Which Color Filter to Choose
The tool offers 7 color shift options: None, Green, Red, Blue, Yellow, Purple, Cyan. The color shift is reapplied at every iteration, so it compounds — more passes means stronger color bias.
Green is the most recognizable "classic deep-fry" style, tracing back to how early memes degraded on low-quality image hosting. Other colors work well for single-tone editorial content.
Batch Processing and Download
Upload up to 20 images (JPG, PNG, HEIC, WebP) and all are processed with the same settings. After processing, download images one at a time or click to get all results in a single ZIP file. The result view includes a slider that lets you drag across the before/after boundary to see exactly how much the image changed.