Cyberduck for Windows doesn't generate correct URL for custom S3 endpoint using <access-key-id>@<namespace-endpoint> format

[jimin-aot]

Cyberduck URL Formatting Issue – Access Key & Path Affecting Hostname

Hi Cyberduck community,

I’m trying to connect to a custom S3-compatible object storage using Cyberduck on Windows, but Cyberduck isn’t constructing the expected URL and fails to authenticate.

---

Expected Behavior

My colleague was able to connect using the following URL format automatically populated by Cyberduck:

https://<access-key-id>@<namespace-endpoint>

I assume providing the bucket name in the Path (e.g. test-media) should only set the folder view—not affect the host.

---

Actual Behavior

In my setup:

1. After entering Access Key ID and server endpoint, Cyberduck built:

   https://<namespace-endpoint>
   

   — the access-key-id@ prefix is missing, and authentication fails.

2. When I set Path to /test-media, the URL changed to:

   https://<namespace-endpoint>/test-media
   

   — causing a DNS resolution failure.

3. I also tried entering:

   https://<access-key-id>@<namespace-endpoint>
   

   manually in the Server field, but it still failed to connect.

---

What I’ve Tried

• Tried using S3 (HTTPS) and Amazon S3 profiles

• Entering the full URL manually with a <bucket-name>@ in the Server filed

• While troubleshooting, I found Cyberduck documentation referencing this:
  Connecting using Deprecated Path Style Requests

  • The documentation mentions a profile and the preference s3.bucket.virtualhost.disable, but the instructions are unclear, the link appears broken, and I couldn’t locate this setting anywhere in the Cyberduck Windows UI. Additionally, I’m not even sure this documentation is relevant to my specific issue.

---

How can I get Cyberduck to prepend access-key-id@ automatically to the hostname?
Any advice would be greatly appreciated—thanks in advance!

[jimin-aot]