Searchable brand asset library for your Amazon S3 files

Build me a brand asset library app so our marketing team can find, preview and share the files we keep in Amazon S3 without logging into AWS or pinging an engineer. The people using it are marketers and designers, not engineers, so it should feel like a media library rather than a file system.

The main screen is a searchable grid of assets from a chosen bucket and key prefix. Use Amazon S3 List Objects (V2) with the prefix to load the objects, paginating with the continuation token since each page returns at most 1,000 keys. Each card shows an image preview plus the file name, file type, size and last modified date. The listing already returns size and last modified for every key, so use those directly and call Head Object for the content type and any richer metadata, fetching it lazily for the cards actually in view rather than calling Head Object for the whole bucket up front.

For previews, fetch the bytes with Get Object in a server side handler and render them in the grid. Images get a real thumbnail. Non image types like PDFs, video, fonts and archives get a clear file type badge instead of a broken preview. Load thumbnails lazily as cards scroll into view so opening the library does not download the entire bucket.

Tags come from Get Object Tagging, using a small deliberate vocabulary like campaign=spring-launch and status=approved. The grid can be filtered by file type, by folder (the key prefix segment) and by tag. Default the grid to status=approved so the safe choice is the first thing anyone sees, with an obvious control to widen the view to everything. Assets tagged status=retired must stay visible rather than being hidden, shown dimmed with a clear Retired badge so people can see the old logo exists and know not to reuse it.

Clicking an asset opens a detail panel with a bigger preview, the full metadata (key, folder, file type, size, last modified, storage class and every tag) and a download button backed by Get Object.

Users can upload new assets straight into the right folder with Put Object. Let them pick the destination folder from the folders already present under the prefix, set the correct content type from the file, and prompt them to tag the asset immediately after the upload succeeds so nothing lands untagged.

Tags are set and edited from the detail panel with Put Object Tagging, including quick actions to mark an asset approved or retired so people stop reusing outdated logos. Two rules matter here. Put Object Tagging replaces the entire tag set, so always read the current tags with Get Object Tagging first, merge the change, and write the full set back, otherwise editing one tag wipes the rest. Amazon S3 also caps objects at 10 tags, so enforce that in the UI with a clear message instead of letting the write fail.

Selecting one or more assets in the grid and hitting Share to Slack opens a composer where the user picks a channel and writes a short description. Populate the channel picker with List Channels from the Slack Bot integration, and post with Slack Bot Send a Message. Send one tidy message listing every selected asset with its name and file type plus the description, rather than one message per file. Use the Slack Bot integration for this, since posts should come from the bot identity. If any selected asset is tagged retired, warn the user before posting.

Remember each user's last used bucket, prefix, filter state and default Slack channel so the library opens where they left off. Keep a lightweight share history recording who shared which asset, when, and to which channel, and surface that asset's history in its detail panel.