clip2file

Save whatever is on the clipboard to a sensibly-named file.

You screenshot something, or copy a chunk of JSON, and now you want it on disk. The usual routine is: open an editor, paste, think of a filename, pick a folder, save. clip2file collapses that into one command — and picks the extension for you by looking at what you actually copied.

$ clip2file
Saved image -> clip_2026-03-04_153045.png  (284.1 KB)

$ clip2file
Saved json -> clip_2026-03-04_153112.json  (1.2 KB)

$ clip2file
Saved text -> clip_2026-03-04_153140_meeting-notes-for-friday.txt  (612 B)

Install

pip install clip2file            # text only, zero dependencies
pip install "clip2file[image]"   # adds image support via Pillow

Python 3.9+.

What it detects

Detected	Extension	How
Image	.png	Clipboard bitmap, via Pillow
JSON	.json	Actually parses it — malformed JSON stays .txt
CSV	.csv	Consistent column count across 2+ lines
URL	.url.txt	A single line that parses as a URL
Code	.txt	Language keyword hints
Text	.txt	Everything else

Filenames are timestamped and, for text, get a short slug taken from the first line — so you can tell your clips apart at a glance without opening them.

Usage

clip2file                    # save to the current directory
clip2file -d ~/screenshots   # save somewhere specific (created if missing)
clip2file -n notes.txt       # use an exact filename
clip2file --text             # ignore images, take text even if a bitmap is there
clip2file --no-describe      # bare timestamp, no content hint in the name
clip2file --print            # show what's on the clipboard, write nothing
clip2file -q                 # print only the path, for piping

Nothing is ever overwritten by accident — a collision becomes note-2.txt, note-3.txt and so on. Pass -f if you actually want to clobber.

Piping works how you'd expect:

code "$(clip2file -q)"       # save the clip and open it

As a library

The clipboard sits behind a small protocol, so the detection and naming logic is usable and testable on its own:

from clip2file import ClipContent, ContentKind, detect_kind, save_clip

detect_kind('{"a": 1}')       # ContentKind.JSON
detect_kind("a,b\n1,2")       # ContentKind.CSV

class MyClipboard:
    def read(self):
        return ClipContent(kind=ContentKind.TEXT, text="hello")

result = save_clip(MyClipboard(), "./out")
print(result.path, result.size_bytes)

Platform notes

Image support needs Pillow on every platform. Text works with no dependencies at all, falling back to the OS clipboard tool:

• Windows — Get-Clipboard
• macOS — pbpaste
• Linux — wl-paste, xclip or xsel

Development

pip install -e ".[dev]"
pytest

The tests fake the clipboard, so they run anywhere — no display server needed.

License

MIT