COMPARISON · MEASURED 2026-09
ImgFlow vs Squoosh: fine-tune one image, or ship a hundred
The short version: neither tool uploads your images — Squoosh is Google Chrome Labs’ open-source project, our live test saw zero third-party requests, and its privacy stance is clean. The choice depends on the job: to fine-tune one image (A/B any two encoder settings, try the experimental WP2/QOI codecs), Squoosh’s compare view goes deeper; to ship a batch (hundreds queued, compressed to 100 KB, exported to social-platform specs, zipped in one click), ImgFlow’s batch queue + target-size engine + grouped scene presets is a shape Squoosh does not have. Both are free; keeping both installed is a perfectly rational setup.
Squoosh https://squoosh.app/
Head-to-head
| ImgFlow | Squoosh | |
|---|---|---|
| Where processing happens | In-browser, 0 upload endpoints | In-browser (open-source and auditable; their own line: “Images never leave your device”) |
| Positioning | Ship batches: queue + ZIP + scenario presets | Fine-tune one image: side-by-side quality slider |
| Batch processing | Queue hundreds + one-click ZIP | No batch queue / ZIP export |
| Target-size compression | Yes (hit exact sizes like 100 KB / 500 KB) | No — you iterate the quality slider by hand |
| Platform presets | 7 scene presets in 3 packs (social, avatar, ecommerce) | None |
| Output formats | JPG / WebP / PNG / AVIF / JPEG XL | AVIF / MozJPEG / OxiPNG / WebP / JXL (beta) / WP2 (unstable) / QOI |
| EXIF / GPS metadata | Stripped by default | Depends on the chosen codec options; no unified default promise |
| Ads & tracking | Skippable slots only, no tracking | Zero ads, zero third-party requests (measured) |
| Interface languages | 5 (EN / 中文 / 日本語 / ES / PT) | Multiple (maintained by the Chrome team) |
| Cost | Core tools free forever | Free and open source |
Where Squoosh wins
- A deeper compare view: each pane picks its own encoder and quality — any A/B pair. ImgFlow’s slider is original vs output only
- Experimental codecs right in the dropdown: WebP v2 (unstable) and QOI are there to play with
- Built by Google Chrome Labs and open source — trust rests on the code, not just promises
Where ImgFlow wins
- A complete batch shape: hundreds queued, failures never block, one-click ZIP — no per-file babysitting
- Target-size engine + grouped scene presets: hitting 100 KB or social/listing/avatar specs is one click
- A dedicated page per format × scenario in 5 languages — searching “HEIC to JPG under 100 KB” lands you exactly there
FAQ
Do Squoosh and ImgFlow both keep images local?
Yes. Both process in-browser: Squoosh’s own tagline is “Images never leave your device”, and ImgFlow has zero upload endpoints. In September 2026 we dropped real images on both and watched the network: Squoosh made zero third-party requests, ImgFlow likewise sent no image traffic. The shared self-test: disconnect the network and keep compressing — if it works, it’s local.
Which one should I use to hit an exact KB size (say 100 KB)?
ImgFlow: the target-size engine converges on the requested size predictably, in one pass. Squoosh lacks the feature — you nudge the quality slider, export, check the size, repeat. Tolerable for one image; unrealistic for a folder.
How we verified
Verified September 2026: live test on squoosh.app (dropping an image, monitoring requests) plus reading the in-app codec list. The project is active; features may change.