An S3 browser built for browsing.

S3 Viewer provides a folder prefix filter, inline previews, downloads, and team sharing across connected S3-compatible providers. Guarded rename, replace, and multi-select deletion all work on AWS S3, R2, B2, and other endpoints — AWS S3 and R2 enforce the safety conditions directly, and B2/other endpoints get the same checks from S3 Viewer immediately before each write.

The AWS console is the right tool for configuring S3 — creating buckets, managing IAM, setting up replication, configuring lifecycle rules. It's not designed around daily file browsing. S3 Viewer focuses on that browsing layer: rename, inline previews, multi-select delete, share links, and every account's buckets in one sidebar. Use both — the AWS console for setup, S3 Viewer for daily work.

Those are desktop apps — they store credentials on your machine, run on one OS at a time, and have no shared-team model. S3 Viewer runs in any browser, encrypts credentials in your browser before they reach the server (or you self-host the whole thing inside your VPC), and ships team access — invite by email, give each member a role, remove them without rotating keys.

Yes, on any connected provider: use ⋮ → Rename. S3 Viewer checks the file you reviewed, preserves readable permissions and settings, and refuses an occupied destination. There's no size cap — files above 5 GB copy in multipart parts automatically. Only an SSE-C encrypted object refuses the action, since S3 Viewer never has the customer-supplied key.

Not yet — search is a prefix filter on the folder you're currently in, and it doesn't recurse or span buckets. For a recursive sweep use aws s3 ls s3://bucket --recursive | grep <query>, or S3 Inventory + Athena on very large buckets.

The preview is picked from the content type your bucket reports for the object, so anything image/*, video/*, audio/*, application/pdf or text/* renders inline, plus JSON, XML, YAML and SQL. That covers PNG, JPEG, GIF, WebP and SVG; MP4, WebM and Ogg; MP3, WAV and FLAC; PDF; and Markdown, CSV, HTML, CSS and source code. Objects stored as application/octet-stream have no content type to go on, so they download instead. Previews are capped at 50 MB.

Yes. Connect any of them to browse, preview, download, upload, rename, and share files, the same as AWS S3. B2 and MinIO's conditional-write enforcement isn't independently verified, so S3 Viewer checks with a HEAD request immediately before each write instead of trusting the endpoint; replacement also needs readable original settings, which some MinIO setups cannot provide, in which case it's refused rather than guessed at. Private endpoints require a trusted self-hosted installation configured to allow them.

Yes — it's free to use and MIT-licensed on GitHub. Self-host it or use the hosted build.