Why did my ZIP fail after my photos converted?

Your photos can convert successfully while their ZIP download still fails. Here's where that handoff can break, and how FluffyBench helps keep it moving.

FluffyBench · 7 min read · Reviewed October 15, 2026

Why can a ZIP fail after the photos convert?

Big photo batches now download on Firefox for Android with FluffyBench Image Space Saver's Download as one ZIP as it runs option enabled and the page kept open until the ZIP finishes.

Making a photo and delivering it are separate jobs, like packing a parcel and carrying it away. In a private or Incognito window, reading a finished photo can fail; on Firefox for Android, the download can time out while it waits for the next photo.

Picked photo, conversion, finished result, then ZIP download. The private-window problem is at reading the new result; the download timeout is at delivery. — Conversion can finish before either handoff fails. How we reviewed

Why can a private window fail to read a finished photo?

A private window has less room to hold the finished copies that the ZIP needs to read. It's like a hallway filling with packed parcels: FluffyBench clears photos into the ZIP as they're ready and limits how much waits.

Without a limit, finished photos pile up while the ZIP falls behind. With FluffyBench's limited queue, fewer photos wait and the ZIP keeps taking them away. — Schematic: limiting the waiting area keeps finished copies from piling up.

When space for finished copies runs out in a private window, conversion can finish yet the ZIP can't read the new copy. In a desktop check, a private window couldn't read a newly created file; a normal window could read one twice as large.

Reading newly created files

Scale: 4.3 GB

Private window: read failed: 2.1 GB
Normal window: read succeeded: 4.3 GB

The picked original stayed readable when the private window couldn't read the new copy.

Why does Firefox on Android stop a ZIP during a big photo?

When we tested Firefox on an Android phone, the ZIP received no new bytes for 30 seconds while a 108-megapixel photo was still processing, and Firefox gave up. FluffyBench sends an empty ZIP block every 5 seconds during the wait, like a tap on a phone line to keep it active.

Recorded failure: no new bytes for 30 seconds, then DownloadFailed. Safeguard: empty blocks at 5 seconds intervals while processing continues. — Top: the download stops while processing continues. Bottom: empty blocks keep bytes arriving without adding photos.

With the empty blocks, downloads kept going through 45- and 75-second pauses, and the ZIPs opened fine.

Do big ZIPs finish in private windows and on phones?

Yes. Large photo batches finished downloading as ZIPs in a desktop private window, Chrome Incognito on Android, and a Firefox private tab on Android.

Finished ZIP entries

Desktop Chromium · private window · 19.1 GB: 9,938 entries
Android Chrome · Incognito · 2.13 GiB: 1,193 entries
Firefox Android · private tab · 380.2 MB: 190 entries

Desktop: 2 h 3 min, 0 checksum errors. Firefox Android: 4 min 36 s. Counts are entries in the completed ZIPs.

Doing it in FluffyBench Image Space Saver

Use FluffyBench Image Space Saver to make smaller photo copies and feed them into one ZIP as they finish. The walkthroughs use a flower photo, a small transparent image and an unreadable JPEG to show how the checks work.

Here's download-as-you-go in FluffyBench's WebP converter: one photo is already in the ZIP while the next is still converting.

Two photo cards in FluffyBench's converter: a pear photo marked In ZIP beside an orange photo still marked Converting. — Keep the page open until the ZIP finishes. · Photo: Rebekah Vos, Unsplash License, via Unsplash. · Photo: Jason Leung, Unsplash License, via Unsplash.

Choose your photos and switch on the download-as-you-go option before processing starts.

Step 1. Choose your photos. Keep your originals until you've checked the saved copies.
1. Choose images opens your file picker.
Step 2. Open Settings. This example keeps the starting photo settings.
1. The gear button opens Settings.
Step 3. Tick Download as one ZIP as it runs: finished photos enter the download as they're ready.
1. Tick the box before processing starts.
Step 4. Tap the Settings gear again to close the panel before processing.
1. The same gear button closes Settings.

Start processing, then check the photo results and that the ZIP finished saving.

Step 1. Press Process 2 images. The count covers the two admitted images; the unreadable JPEG was rejected.
1. Process 2 images starts processing and the ZIP download.
Step 2. Check Done and Needs a look when processing ends.
1. Done counts images that finished processing.
2. Needs a look includes warnings as well as failures.
Step 3. In ZIP means a photo entered the download; Failed means a file was left out.
1. In ZIP means this photo entered the download.
2. A warning can appear on a photo sent to the ZIP.
3. Failed means this file did not enter the ZIP.
Step 4. Check that the ZIP finished in your browser's downloads list. Shown here: desktop Chromium, where Show in folder is available for the saved ZIP. Firefox on Android has a different Downloads screen.
1. The completed download lists the ZIP by name.
2. Show in folder is available for the saved ZIP.

Converting to one format instead? FluffyBench's WebP converter has the same Download as one ZIP as it runs box once you've added at least two images, so finished WebP conversions can go straight into a ZIP.

Use FluffyBench Image Space Saver to make smaller copies and download them as one ZIP

Image Space Saver

How we reviewed: storage checks and completed ZIPs

We checked delivery of large archives by running photo batches in desktop and phone private windows. The table separates selected files from ZIP entries and records exclusions; the device and browser details are below.

Run	ZIP and time	Inputs and checks
Desktop Chromium · private window	19.1 GB · 2 h 3 min	9,938 entries; 0 checksum errors; 62 files left out as image sequences.
Android Chrome · Incognito	2.13 GiB (binary gigabytes) · 31 min	1,200 files selected; 1,193 entries in the ZIP. Excluded selections were not converted.
Firefox Android · private tab	380.2 MB · 4 min 36 s	201 files picked; 10 HEIC files and a manifest left out. Archive integrity, headers, duplicate names, and file names checked out.

How we reviewed

Reviewed October 15, 2026

FluffyBench checks of private-window storage, temporary copies of picked files, and ZIP downloads. Each note lists the browser and device used.

The historical probes were not repeated for this review. Phone timings have no repeat distribution, and the private-storage summary does not give a complete run count for every browser. GB uses decimal bytes; GiB and MiB use binary units. The historical evidence supplies dates and browser versions, not a product build id for each probe. Results describe these devices, browsers, and inputs; they do not establish a private-window size limit or promise the same result on other devices.

Private-window read check, 2026-10-06: desktop Chromium allocated new Blob data (file-like data held by the browser) in a private window and in a normal browser profile, then tried full reads, reads in small pieces, reads of the picked file, release, and forced garbage collection. Releasing the data and collecting it restored private reads in that check. A separate Android check used a OnePlus 6T with Android 10 and Chrome 153 in a private window; reads succeeded at 79,691,776 bytes and failed at 80,740,352 bytes. Those are two observed sizes on that device and build.

Storage-limit check, 2026-10-07: bundled Chromium 153 and installed Chrome 155 in private desktop windows wrote temporary files until a quota error, repeated the process, and cleared storage between checks. Observed write capacity was 7.2 GB against a browser-reported estimate of 17.9 GB.

Firefox Android failure, 2026-10-07: OnePlus 6T, Android 10, Firefox 156. A small photo was followed by the large JPEG (108 MP) after the ZIP had started; converting without a ZIP succeeded. 185,526 bytes had arrived before the flat interval, which ended 30,000 ms later with DownloadFailed; the source gives no earlier arrival time for those bytes. The browser investigation found a 30,000 ms default read timeout. Firefox's retry reached an already claimed or missing download stream and returned HTTP 409/404. The writer's default time between empty blocks is 5,000 ms, and the tests that recreated only the long pause used one download request (one GET).

Separate pause checks waited 45 s and 1 min 15 s between image entries and produced valid ZIPs. The batch runs establish delivered archives; the pause checks establish the empty-block safeguard separately. The live-ZIP illustration is an existing real converter capture; the preparation and finish walkthroughs show Image Space Saver.

Storage policy: the finished-result queue uses a budget of 256 MiB, or 0.5% of reported device memory when that is smaller. An empty queue accepts one result exceeding that budget so work can continue. ZIP data ready to send uses the same budget value. On Android, temporary copies of picked files are limited to 128 MiB across a batch or 64 MiB for any one file. Image-processing data and bytes already sent to a download have separate lifetimes; the source reports no measured success rate.

Desktop run: 1 historical run on Windows 11 with Chromium 153 in a private window, using repeated generated test images and checksum, name, and header checks on every entry. The archive uses the ZIP64 format. The source records the 62 left-out files as inputs the tool turned away as image sequences instead of converting them.

Chrome phone run: 1 historical run on a OnePlus 6T with Android 10 and Chrome 153 in Incognito. Files that were left out were not converted.

Firefox phone run: 1 run on a OnePlus 6T with Android 10 and Firefox 156 in a private tab, using the native file picker with Space Saver WebP quality 80 and resize/target off; full archive integrity, header, duplicate-name, and file-name checks passed.