Downloading
An Assembly can return temporary download URLs for the
processing results it includes.
Find these URLs in the results of the Assembly Status JSON.
By default, results are included for Steps that no other Step uses. Set "result": true to also
include an intermediate Step's results.
Download temporary results to your own storage as soon as the Assembly completes, and within 24 hours. Result URLs stop working after the temporary files expire. Whether a processing result is stored also depends on the temporary-file retention policy of its Template or Workspace. Assembly Status JSON retention is a separate setting and does not extend temporary-file storage.
For production workflows, we recommend an export Robot that
saves results to your cloud bucket or server. Exports are retried on failure; by default, if the
export still fails after retries, the Assembly fails rather than reporting a successful export.
Access to files in your own storage is controlled by that provider and your configuration. Check the export Robot's
acl default: some exports are public unless you configure private access. The
Google Cloud Storage guide and
DigitalOcean Spaces guide show private setups.
Temporary URLs (tmp-*.transloadit.net) are strictly for downloading files once to your own storage:
- ✅ Allowed: Download the file once and store it on your infrastructure
- ❌ NOT Allowed: Share temporary URLs with your end users (violates Terms of Service)
- ❌ NOT Allowed: Embed temporary URLs in your application/website
- ❌ NOT Allowed: Use as a CDN or for repeated downloads
Violations of these restrictions breach our Terms of Service and may result in immediate account suspension. Files accessed more than 10 times or consuming excessive bandwidth are prohibited. Always use an export Robot to save files to your own storage for production use.
🤖/file/filter passes accepted files on to subsequent Steps and does not
emit rejected files as its results. It reuses the input file's URL information instead of creating
another temporary stored copy. By default, uploads have temporary storage; filtering an upload
reuses that existing copy without extending its lifetime. Output from a processing Step that feeds
the filter is not kept in temporary storage by default, so a working URL alone does not establish
how long that output will remain available. To request temporary storage, set "result": true on the producing Step. This
also includes that Step's unfiltered results in the Assembly Status JSON and remains subject to
your retention policy. To save only accepted files to your own storage, connect an export Step to
the filter Step instead.
For delivery through Smart CDN, follow the Smart CDN setup guide. Temporary download URLs are not a substitute for Smart CDN or your own hosting.