Step 1 / source
Choose the source
Drop one file here
Choose a file when a ZIP container is acceptable to the recipient and you want to test whether compression helps.
Checking publication status...LOCAL COMPRESSION CHECK / FILE TO ZIP
Try lossless ZIP compression on one file and offer a download only when the resulting archive is actually smaller.
Step 1 / source
Choose a file when a ZIP container is acceptable to the recipient and you want to test whether compression helps.
Checking publication status...This operation runs with browser APIs. Nirmion receives catalogue requests, not the selected source bytes.
The converter creates a separately named output. It never writes back into the source file.
One file up to 100 MB. The result is a ZIP containing the byte-identical source; already compressed formats may produce no downloadable result.
Plain-language method
Nirmion applies lossless DEFLATE compression inside a ZIP, compares the complete archive size with the original, and exposes the ZIP only when it saves bytes. The file stored inside remains byte-for-byte unchanged.
Text, CSV, JSON, XML, and some raw data often compress well. JPEG, PNG, MP4, PDF, DOCX, and existing archives are commonly already compressed.
Fast uses less effort, balanced suits most files, and smallest tries the highest ZIP compression level. All three preserve source bytes.
A useful result ends in _nirmion_tools_reduced.zip. If ZIP headers or incompressible data make it larger, the page reports zero saved and provides no misleading download.
Useful when
Before converting
Know the boundary
The tool does not lower image quality, video bitrate, PDF resolution, or document fidelity. It only applies lossless ZIP compression and may find no saving.
Quick answers
An honest no-gain result is useful information.
The attempted ZIP was the same size or larger, so offering it would not reduce the file.
No. Successful ZIP extraction returns the exact original bytes.
Balanced is a practical default; smallest may help compressible data but takes more work.