ZIP packages multiple entries with optional per-entry compression.
TAR.XZ applies XZ compression to a sequential TAR archive.
Regular files and directories are repackaged without changing their file bytes.
Links, device nodes, duplicate paths, unsafe paths, and archives beyond the entry or expanded-size limits are rejected.
The worker repackages regular files and directories. It rejects links, devices, duplicate paths, unsafe paths, more than 200 entries, more than 100 MB expanded data, and unsafe ZIP compression ratios.
One archive is processed per request. The input limit is 100 MB.
The web application validates the file before forwarding it to a private worker. The worker validates the signature again, processes the file inside a unique temporary directory, validates the output, and removes the job directory in a guaranteed cleanup path. Files are not written to the database or retained for analytics.
When you import from a URL, the application fetches only a validated public HTTP or HTTPS address, checks the downloaded file in memory, and applies the same worker cleanup rules. The URL, filename, and file contents are not sent to analytics.