Files Brought in by Link
Part of the MountDev Ops Files documentation. Updated November 19, 2026
MountDev Ops Files documentation | Updated November 19, 2026
An outside service your company works with can hand your Files a short-lived download link for a file, and your workspace fetches that file straight into your own Files. The link is used once and is never kept. It is not a screen: nobody clicks anything, and the file appears in your Files like any other file brought in from elsewhere.
How It Works
The service connects to your workspace with your company's own engine connection and sends one call for each file: the link, the file's size, its SHA-256 checksum, a name, and, if wanted, the folders it belongs in and the records it is attached to. Your workspace then:
- Checks that the address is safe: it must start with
https://, use the standard port, and point to a public address. A private, local, or internal address is refused, and so is a redirect to one. Each redirect is checked the same way, and a link that keeps redirecting is stopped after a few hops. - Fetches the file once, a piece at a time, so the whole file is never held in memory.
- Checks the size and, when it was given, the checksum of the whole file. A file that does not match is refused and nothing is stored.
- Encrypts the file with your company's own key and stores it in your own Files, the same way every uploaded file is stored. It counts as File storage on Usage.
A file with the same id and checksum that is already in Files is never brought a second time, and its link is not even fetched. A file you deleted in Files is never brought back.
What Is Never Kept
The link is used in memory for that one fetch. It is never stored in your workspace, never written to a log, never put in your Activity Log, and never sent back in an answer or an error message. If a link has expired or was refused, the answer says only that the link was refused or has expired. The sender then tries again with a fresh link.
Where It Lands
The file lands in a folder tree the sender names, under From Hub Archive, and is attached to the records the sender names, with no copy. It is quiet: no events, webhooks, automations, notices, or emails run, and each file is one line in your Activity Log with the file's name, never the link.
Limits
- One link call brings in one file of up to 2 GB. A larger file is sent in pieces instead.
- The call waits (an answer asking the sender to try again in a few seconds) while someone is using the workspace, like every other import.
- Your workspace answers only for itself. A call made to another company's workspace never reaches your Files, and the reverse.
Who Can Use It
No person, API key, or AI agent can use it. It is a system call for your company's own engine connection only.
