PART File Documentation

---

Technical details

Feature	Value
Full name	Partial Download File
File extension	.part (appended, e.g. movie.mp4.part)
MIME type	application/octet-stream (no type of its own)
Category	Web / temporary file
Format type	Incomplete copy of the file being downloaded
Has its own structure	No — contents match the target file so far
Magic number	None (header is whatever the real file's header is)
Created by	Mozilla Firefox; wget, aria2, KGet, JDownloader
Typical location	Your Downloads folder, beside the target file
Companion files	Sometimes a control sidecar (e.g. aria2's .aria2)
Resumable	Only if the downloader kept the session and the server supports HTTP range requests
Temporary	Yes — renamed to the real name on completion
Safe to delete	Yes, if no download is currently using it
Related extensions	.crdownload, .partial, .download, .opdownload, .!ut

What is a PART file?

A .part file is the staging copy a download writes while data is still arriving. It is not a file format. When you download something in Mozilla Firefox, the browser opens a file named after the target with .part appended — for example ubuntu.iso.part — and streams the incoming bytes into it. Only when the transfer completes does Firefox rename it to the real name, ubuntu.iso. The .part suffix exists precisely so that a half-finished file is never mistaken for a usable one, and so the operating system will not treat an incomplete installer as runnable.

Firefox is the best-known producer, but the convention is shared by many tools: the command-line downloaders wget (with --continue) and aria2, the KDE downloader KGet, and managers such as JDownloader. It is the Firefox-world equivalent of Chrome and Edge's .crdownload, Safari's .download, and old Edge's .partial — all the same idea under different names.

What is actually inside a .part file

There is no wrapper, header or encryption added by the downloader. A .part file holds the leading bytes of the real file, in order, exactly as they came off the network. A partially downloaded MP4 therefore starts with the normal MP4 box header; a partial ZIP starts with the ZIP local-file signature. That is why .part has no magic number of its own: whatever the target file's first bytes are, those are the first bytes of the .part. A 0-byte .part means the transfer was requested but never really began.

Some tools keep a small companion control file next to the data so an interrupted job knows which byte ranges it already has. aria2 writes a .aria2 sidecar for this; older Firefox builds used a .part.json style record. That control file, not the .part itself, is what lets a multi-connection downloader resume in the middle rather than from the start.

Finishing or recovering the download

Because a .part is only unfinished data, the useful question is never "how do I open it" but "how do I finish it". In Firefox, press Ctrl+J to open the Downloads panel, find the interrupted item, and click the retry or resume arrow. On the command line, re-run with the resume flag: wget -c URL or aria2c -c URL continues into the existing .part instead of starting a new file. JDownloader and other managers resume from their queue.

Resuming only works when two conditions hold: the downloader still has the session (you have not cleared the download history), and the server supports HTTP range requests. If either fails, the partial usually cannot be continued and the download must be restarted from the beginning. Renaming a .part to the real extension recovers a working file only when the download had effectively reached 100% and just failed to rename — compare the file size against the expected size. A half-downloaded archive, installer or document stays corrupt regardless of what you call it, because renaming changes the name and nothing else.

The one partial-success case: media files

Most incomplete files are useless until they finish, but timed media is the exception. A video or audio container is roughly linear, so VLC can often play the portion of an incomplete MP4 or MKV that has already arrived. Drag the .part into VLC and it will play up to the point the download reached, then stop. It cannot invent the missing bytes, and seeking past the downloaded region will fail, but for checking whether a large video is the right one this often works.

Is it safe to delete, and is it dangerous?

A .part cannot execute. It is exactly as risky as the file being downloaded, no more and no less, and the suffix actively helps by stopping the OS from running an incomplete executable. Deleting a .part is safe whenever no download is currently using it; doing so simply gives up any chance of resuming that particular transfer, so resume first if you still want the file. Orphaned .part files left behind after a cancelled or finished download are safe to remove and only waste disk space. One caution: if you did not start the download that created a .part, treat the underlying source as suspicious and delete it rather than completing it.

Frequently asked questions

Can I just rename a .part file to the real extension?

Only if the download actually completed. Renaming changes the filename, not the data, so an incomplete file stays broken. Check the size first: if it matches the expected download size, removing the .part suffix should give you a working file; if it is smaller, resume or re-download instead.

Why is my download stuck as a .part file?

The transfer did not finish or the final rename failed. Common causes are a dropped connection, the browser closing mid-download, an antivirus scan holding the file, or a disk-space or permission problem. If the size matches the target, rename away the .part; otherwise resume the download or start it again.

References

• Mozilla Support — find and manage downloaded files in Firefox
• GNU Wget manual — resuming downloads with --continue
• aria2 — multi-protocol download utility

Feedback

Was this page helpful?

Yes No

Glad to hear it! Please tell us how we can improve.

Sorry to hear that. Please tell us how we can improve.