Amazon S3 has no zip command. Neither the AWS console nor the CLI can turn a folder of objects into a single archive by itself. So how do people actually do it? There are four practical routes, and they differ a lot in effort and cost.
1. Console: download, zip, upload
Select the folder in the console, download the objects, zip them on your computer, and upload the archive back. Fine for a handful of small files. Painful for anything large: you download every file and pay egress for it, and the process is manual.
2. AWS CLI: copy locally, then zip
aws s3 cp s3://my-bucket/reports/ ./reports/ --recursive
cd reports && zip -r ../reports.zip .
aws s3 cp ../reports.zip s3://my-bucket/reports.zipScriptable, but it needs local disk for the whole folder and downloads every byte. For a 50 GB folder, you pay for 100 GB of transfer and need 50 GB of disk.
3. EC2 or Lambda: stream the folder through a zip writer
Run code that reads each object and writes it into a zip stream, uploading the archive back to S3. This avoids local storage, but you build and operate the code. Lambda works for small and medium sets but has a 15-minute limit and limited temporary storage; large folders need EC2 or a queue-based worker.
4. An API that does it for you
S3zipper takes a folder prefix, expands it, compresses the files inside AWS, and writes the archive back to your bucket. One HTTP request, no server to run:
curl -X POST https://api.s3zipper.com/v1/zipstart \
-H "Authorization: Bearer YOUR_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"awsKey": "AKIA...",
"awsSecret": "...",
"awsBucket": "my-bucket",
"awsRegion": "us-east-1",
"filePaths": ["my-bucket/reports/2026"],
"zipFileName": "reports-2026.zip"
}'You get a task id back immediately, poll /v1/zipstate, and fetch the signed download URL from /v1/zipresult.
Which route should you use?
- A few tiny files, one time — the console is fine.
- Regular automation — an API keeps your code short and avoids moving data twice.
- Very custom packaging logic — a small Lambda or EC2 worker makes sense if you need to transform files first.
Read the full request/response walkthrough in How to start a zip job, or see what works and what doesn't when zipping a whole bucket.