Routes

A route links a domain + path prefix to a bucket + folder. Create it from Routes → New route choosing among verified domains and existing buckets.

Active routes, with the row expanded · 26/09/2026

How to read it

Field	Meaning
Domain	Must be verified.
Path prefix	/ for the whole domain, or /docs/. The panel normalizes it (starts and ends with /).
Bucket	The bucket to read from.
Folder	Optional: everything outside stays private.

The preview in the window shows the match: img.company.com/file.jpg → catalog/photos/file.jpg.

Examples

Route	Request	File read
img.company.com + / → catalog / photos/	img.company.com/boat.jpg	catalog/photos/boat.jpg
media.company.com + /docs/ → documents / public/	media.company.com/docs/prices.pdf	documents/public/prices.pdf
media.company.com + /photos/ → catalog / photos/	media.company.com/photos/boat.jpg	catalog/photos/boat.jpg

The same domain can have several routes with different prefixes, even to different buckets.

Selection rules

• The longest prefix wins, compared by whole segments: /docs/ does not catch /docsx/.
• A domain + prefix pair is unique: a duplicate is rejected with the name of the rule that holds it.
• An unknown host answers 404.
• Only GET and HEAD are allowed (other methods: 405).
• The path is normalized; ./.. and %2F are rejected with 400.
• The query string is not forwarded to the storage and is not part of the cache key.
• The route prefix is removed before looking for the file (in the price list example, /docs/ is not part of the file name in the bucket).

Changes apply immediately, with no restart. To delete a domain or a bucket you must first delete the routes that use them.