S3 Viewer
Open Viewer

How to search across S3 buckets

TL;DR

Amazon S3 has no native search API — every tool lists keys with ListObjectsV2 and filters client-side. The trick is paginating through the full result set instead of stopping at the first 1,000-key page. For small to medium buckets, `aws s3 ls --recursive | grep` is fast enough. For interactive cross-bucket search across providers, S3 Viewer paginates in the background with Tab autocomplete. For buckets with millions of keys, S3 Inventory + Athena is the right pattern.

Steps

Step-by-step.

  1. 01

    In S3 Viewer: type in the search bar

    Search runs as you type. Hit Tab to autocomplete the next prefix segment — finding data/models/v2/training/checkpoint-42/weights.pt becomes a few Tab presses. S3 Viewer paginates through ListObjectsV2 in the background and filters in-memory across the buckets you have open — including buckets on different providers.
  2. 02

    Open the result

    Hits show the full key path so you can see exactly where each match lives. Click to jump straight to the file.
  3. 03

    AWS CLI: recursive list + grep

    Quick and dirty. Works for small to medium buckets.
    aws s3 ls s3://my-bucket --recursive | grep <query>
  4. 04

    AWS CLI: list-objects-v2 with --prefix

    Use --prefix to narrow the listing first, then filter the output. Faster on large buckets — listing happens server-side.
    aws s3api list-objects-v2 \
  --bucket my-bucket \
  --prefix logs/2026/ \
  --query 'Contents[?Size > `1000000`]'
  5. 05

    For very large buckets: S3 Inventory + Athena

    Enable S3 Inventory for a daily CSV/Parquet manifest of every key, then query it with Athena. The right pattern for buckets with millions or billions of objects — listing them via the API would take hours.
Under the hood

What's actually happening.

S3 has no search API — you list keys (ListObjectsV2) and filter client-side. The AWS console's filter operates on the visible page (1,000 keys), which is fine for small buckets but not for deep ones. S3 Viewer paginates through ListObjectsV2 in the background and filters in-memory; results stream in as pages come back. Tab autocompletes the next prefix segment as you type, so navigating deep paths is keyboard-fast. The same pattern works for Cloudflare R2, MinIO, Backblaze B2, and any other S3-compatible provider, since they all implement the same listing API. For multi-million-key buckets, S3 Inventory + Athena is the right pattern.

FAQ

Common questions.

How do I search for a file in an S3 bucket?

S3 has no search API — you list keys with ListObjectsV2 and filter client-side. From the CLI: `aws s3 ls s3://bucket --recursive | grep <query>`. From a tool: open the bucket in S3 Viewer and type in the search bar — it paginates ListObjectsV2 and filters as you type, with Tab autocomplete for prefix paths, across the entire bucket. For buckets with millions of objects, use S3 Inventory + Athena.

Why doesn't the AWS S3 console search find my file?

The console's search filters the keys currently visible on the page (typically 1,000 keys). For a bucket-wide search you need a tool that paginates through ListObjectsV2 and applies the filter to every returned key. S3 Viewer does the pagination in the background, so a query in a 50,000-key bucket finds the file even when it's deep in the result set.

Can I search across multiple S3 buckets at once?

Yes. S3 Viewer's search runs across every bucket you have open in the workspace, including buckets on different providers like Cloudflare R2 and Backblaze B2. From the CLI, you'd loop over a list of buckets and grep each one — there's no native multi-bucket search API.

How do I search a Cloudflare R2 bucket?

R2 implements ListObjectsV2 and works the same way as S3. From the AWS CLI with `--endpoint-url`: `aws s3 ls --endpoint-url https://<account>.r2.cloudflarestorage.com --recursive | grep <query>`. From S3 Viewer, just type in the search bar — R2 buckets are searched the same as S3 buckets, with Tab autocomplete and bucket-wide pagination.

Is there an S3 search API?

No — S3's only listing primitive is ListObjectsV2 with prefix and delimiter. For real search (full-text, by metadata, by content), build it on top of S3 Inventory + Athena, or index the bucket into a search engine like OpenSearch.
Use S3 Viewer for this

Skip the CLI. Try it in the browser.

S3 Viewer turns the steps above into a single click. Open source, self-hostable, free for personal use.

Open ViewerStar on GitHub