Google Takeout Split Into Many Zip Parts? How to Unpack Them
Why Google Takeout splits your photos into takeout-001, -002 and so on, why one album lands in several parts, and how to extract them together.
Your Takeout download is not one file. It is takeout-20260901T101500Z-001.zip, -002.zip, -003.zip and so on, sometimes dozens of them. Unzip one and you find a piece of an album. The rest of that album is in some other part.
This is normal, and it is fixable in a few minutes if you know the rule.
Why Takeout splits the export
When you request an export, Google asks how big each file may be: 2 GB, 4 GB, 10 GB or 50 GB. A 300 GB library at 10 GB per file becomes thirty zips. Google fills each one until it hits the limit, then starts the next. It does not care about album boundaries. An album that straddles the limit ends up half in part 7 and half in part 8.
Every part has the same folder layout inside: Takeout/Google Photos/<album or year folder>/. So Photos from 2019 appears in as many parts as it takes to hold that year's photos. The .json sidecar for a photo is usually in the same part as the photo, but not always. Album metadata.json files may be repeated in several parts.
These are ordinary zips, not a multi-volume archive. You do not need all parts present to open one. Each opens on its own.
Extract them all into the same folder
The fix is to unpack every part into one parent folder and let the folders merge. On a Mac, select all the zips in Finder and double-click; Archive Utility puts each into a Takeout folder in the same place and merges them. If it creates Takeout, Takeout 2, Takeout 3 instead, drag the contents of each Google Photos folder into the first one and choose "Merge" when asked.
On Windows, the built-in extractor handles one zip at a time and is slow with tens of thousands of small files. 7-Zip is free and better. Select all the parts, right-click, choose 7-Zip, then "Extract here". Since each contains a Takeout folder, they merge automatically. Say yes to overwriting duplicate album metadata.json files; they are identical.
On Linux, from the folder with the zips:
for f in takeout-*.zip; do unzip -o -q "$f" -d .; doneThe -o flag overwrites the repeated album files without asking.
Make sure you have room first. The unpacked folder is about the size of all the zips together, so you need double the download size free while both exist. See how much disk space a Takeout needs.
Check that nothing is missing
Downloads fail. A part that ends at 3.98 GB when the others are 10 GB is probably truncated. The zip will either refuse to open or open with errors. Count your parts against the number shown on the Takeout page. Google names parts sequentially, so a gap between -011 and -013 means -012 never arrived.
Takeout downloads expire after seven days. If a part is corrupt and the week has passed, you have to request the export again from the start. It is worth checking on day one.
After extraction, look for photos whose sidecar is missing, or sidecars whose photo is missing. A quick count of images versus .json files in each folder will show a gross mismatch. A small difference is normal; a folder with 900 photos and 12 sidecars means a part did not unpack.
The tgz option
Takeout also offers .tgz instead of .zip. It behaves the same way, split at the size you chose, and it preserves file dates a little better. On Windows you still need 7-Zip to open it. The zips are fine for most people; if you are on Linux or a Mac, .tgz is slightly less hassle.
Working from the zips directly
You do not have to unpack the parts by hand if a tool can read them for you.
Takeout JSON Metadata Fixer takes the zips as input on macOS, Windows and Linux.
Pick one part and it looks for the sibling parts in the same folder and adds them. If a number is missing from the sequence, it tells you before it starts, so you can re-download the part while the link still works. It also checks the free space on the destination drive against the size of the archives and stops early if you will run out halfway.
Then it reads every part as one library, pairs each photo with its sidecar even when the two ended up in different parts, and writes the real date and GPS into each file. You choose whether to keep the original names and folders or to sort everything into year and month folders with readable names. The first 500 photos are free.
If you prefer to unpack by hand and run a free tool such as GooglePhotosTakeoutHelper or exiftool afterwards, that works too. The steps above get you to a single merged folder either way.
Frequently asked questions
Do I have to extract the parts in order?
No. They are independent zips with overlapping folder names. Extract them in any order into the same parent folder.
Why are there more files in Takeout than photos in Google Photos?
Photos in several albums are copied into each album folder, and every photo has a sidecar. Once you remove the sidecars and album duplicates, the count matches. See Takeout duplicates and edited versions.
Can I choose a bigger part size to avoid this?
Yes, up to 50 GB per part. Fewer, larger parts are easier to manage but a single failed download costs you more. 10 GB is a reasonable middle.
One part will not open at all. What now?
Re-download it from the Takeout page if the export is less than seven days old. If it has expired, request a new export. You can limit the new export to Google Photos only to make it faster.