Skip to main content

Grab a video from the web, cut the part you want, caption it, and export a clean GIF or MP4. Everything runs on your PC. No watermark, no account, no upload.

Animated tour of the app


What it does

  • Download from the web. YouTube, Instagram, X, Reddit, Bluesky, Facebook, and 1,000+ other sites yt-dlp supports. Paste a link, press Ctrl+V from anywhere, or queue a batch. Reads browser cookies for private and age-gated videos.
  • Trim to the exact moment. Frame-accurate timeline with a waveform and thumbnail strip. Queue several cuts from one video, then export them separately or stitched with transitions.
  • Captions that look right. Text with outline, shadow, bold, italic, and multiple lines, and the preview matches the exported frame exactly. Auto-caption speech with Whisper, or import an .srt or .vtt.
  • Captions that follow the action. Pin a caption to a moving subject and it tracks them across the frame, the way the "click for more" memes do. Drag to set the path by hand, or choose Follow an object from here and let it follow the subject for you. Preview and export stay in sync.
  • Overlays. Logos, watermarks, and sticker or GIF overlays dragged anywhere on the frame, each with its own size, opacity, and timing. Paste an image straight from the clipboard.
  • Export for the platform. Tune GIFs (dither, palette, loop) or export hardware-encoded MP4s. Reframe 16:9 to 9:16 for Shorts, Reels, and TikTok with a blurred-background fill. Speed, rotate, mute, audio-only, and colour adjustment are built in.
  • Record your screen straight into the editor.
  • Save real projects. .vidkid project files preserve the source link, trim ranges, crop, captions, overlays, and export choices. Autosave recovery protects work after an interrupted session, and recent projects stay one click away.
  • Updates that fit the install. Store installs update through the Microsoft Store; Setup, winget and portable installs open the GitHub release; pip, APT, AppImage, macOS and source installs each get the route that fits them.
  • Runs fully offline. No upload, no account, no watermark. Open source, FFmpeg included.

The app is three steps across the top of the window: Import brings a clip in (a file, a screen recording, or a link), Edit is where you trim, caption, crop and color it, and Export makes the MP4, GIF or MP3. Edit fits on one screen with no scrolling. VideoKidnapper also has undo and redo, project save and recovery, batch export, an export history, keyboard shortcuts (Ctrl+S project, Space play, J/L step, I/O in-out, Q save range, Ctrl+E export; ⌘ on a Mac), five themes, a CLI mode, and a plugin API. See docs/ROADMAP.md for what is next.


Workspaces & features

Edit

The Edit workspace with a video loaded, a caption, and two saved ranges

The media panel sits on the left (collapse it to a thin rail for a bigger preview), the preview with play controls in the middle, the Inspector on the right, and a timeline underneath. The timeline stacks saved ranges, caption clips, image-overlay clips, a filmstrip, and the waveform under one playhead: click to seek, drag the blue in/out edges to pick the part, drag a caption or overlay to move it, drag its edge to retime it, and Ctrl+wheel to zoom. Press Q to save the selection as a range; each range exports as its own clip, or they join into one video with a cut, crossfade, or fade.

The Inspector edits one thing at a time. Clip has the in/out times, speed, rotate, frame shape (center-crop or blurred fill), hand crop, and mute. Text edits the selected caption: multiline text, style preset, font, size, bold, italic, color, outline, shadow, box, position, timing, and fade. Image edits the selected overlay. Color has brightness, contrast, saturation, and gamma, and the preview shows the change so what you see is what you export. Captions render on the preview exactly as they will export. Captions ▸ Import SRT or VTT and Captions from speech (Whisper) both add caption clips to the timeline.

Turn on Record a motion path and a caption follows a moving subject across the frame, like the "click for more" memes. Drag the caption at a few points in the clip to set its path by hand, or drop the caption on your subject and choose Follow an object from here to have it follow them automatically (OpenCV object tracking, bundled in the Store, Windows, macOS, and Linux AppImage builds). The same interpolation drives the preview and the export, so what you see is what renders.

One caption tracked across three moments of a clip

Import

The Import workspace with a pasted link and the batch download list

Import is where a clip comes in: paste a link, open a file, record the screen, or reopen a recent project. Whatever loads lands in Edit. The Kidnap from bar detects the platform as you paste (or press Ctrl+V to drop a link in from anywhere in the app). Supported with brand chips:

Platform Host patterns
YouTube youtube.com, youtu.be, music.youtube.com, m.youtube.com
Instagram instagram.com
Bluesky bsky.app, bsky.social
Twitter/X twitter.com, x.com, mobile.twitter.com
Reddit reddit.com, redd.it, v.redd.it (gallery-wrapped + video+audio auto-merged)
Facebook facebook.com, fb.watch, fb.com, m.facebook.com

…plus the 1,000+ other sites yt-dlp supports. Cookies from reads login cookies from Chrome / Firefox / Edge / Brave / Opera, or a cookies.txt export, for private/age-gated videos. (Windows Chrome encrypts its cookie DB, so close Chrome fully, use Firefox, or a cookies file; the in-app error explains it.) Downloads retry transient failures with resume, and Update compatibility keeps the extractor current. Batch Download takes a list of links, grabs them in order, and loads any finished one into the editor with Use.

Export

The Export workspace with platform presets, format, and quality

This project puts every export decision on one page: what to export (the selection plus saved ranges, optionally joined), what it's made for (YouTube, Shorts, Reels, TikTok, X, Bluesky, Discord size limits, Slack GIF, and more), format, quality, and where it's saved. Quick export in the header (Ctrl+E) exports with these settings from anywhere.

Batch files exports whole local files with one set of settings: a queue where each row can target a different platform, with quality, MP4 or MP3, and the save folder on the side. Speed, rotate, frame shape, and color come from Edit, so there's one place to change them.

History

Export history with thumbnails

Every successful export, from Edit or Batch, is persisted to ~/.videokidnapper_settings.json. The 25 most recent show under Export ▸ History, grouped by day, with a thumbnail, format, quality preset, time, and file size. Play opens the file in the default player; Show in folder opens its folder in Explorer / Finder. Missing files (moved or deleted) are dimmed and their buttons disabled.

Projects and recovery

Project save and recent-file hub

Press Ctrl+S to save a .vidkid project, Ctrl+Shift+S to save a copy, or Ctrl+Shift+O to open one. Project files keep the source video reference, trim and queued ranges, crop, text and image layers, plus export choices. Relative media paths travel with a project folder when possible. Unsaved edits are written atomically to a recovery file, and the next launch offers to recover it after an interrupted session.

App updates

Install-aware update prompt

When a release is available, an Update chip appears in the header (or use Check for updates in the settings menu). Microsoft Store installs update through the Store. Setup and winget installs open the GitHub release, whose Setup installer updates them in place and keeps your settings. pip and APT installs use their package manager. Portable, AppImage, macOS, and source installs open the matching release instead of replacing a running executable in place.

Debug log

Debug log with color-coded levels

Open it from Debug log in the settings (gear) menu.

Captures stdout + stderr with level-colored tags: INFO accent-blue, WARN amber, ERROR red. Uncaught exceptions from Tk callbacks and normal Python code both land here via a global exception hook, so a crash leaves a traceback instead of killing the app. Useful when yt-dlp reports a protected video or ffmpeg rejects a filter chain — the actual error text is here instead of the generic "failed" toast.

Setup

First-run welcome

New installs start with a short three-step welcome and direct actions for opening a local video or using a web link. Editing stays on the computer, with no account or upload required.

Setup dialog

Opened from Setup and components… in the settings (gear) menu. When FFmpeg is missing on first run, a setup screen explains the source, destination, and integrity check before asking permission to install. Each row describes a prerequisite and the feature it unlocks; required items are pre-checked, optional ones wait for opt-in. Select all missing toggles every installable row.

  • Install Selected runs in a background thread: FFmpeg is pulled as a portable build (gyan.dev on Windows, checked against the publisher's SHA-256 digest; on macOS the same static build the Mac app bundles, checked against digests pinned in the source), staged, and then placed in assets/ffmpeg/bin/ (the app's fallback lookup path). Python packages use python -m pip install --user; no admin access is needed.
  • Open Admin Terminal launches an elevated shell pre-populated with the right commands for your OS: winget install Gyan.FFmpeg on Windows (via PowerShell Start-Process -Verb RunAs), brew install ffmpeg on macOS (via Terminal + osascript), sudo apt-get install ffmpeg on Linux. If no terminal is available, commands are copied to the clipboard as a fallback.
  • Relaunch restarts the current process so newly-installed prerequisites are picked up.

Share (inside the Export dialog)

After a successful export, the Export dialog reveals a share panel with a caption entry and one button per supported platform. Clicking a button:

  1. Copies the exported file to the OS clipboard (Windows: PowerShell Set-Clipboard -Path; macOS: pbcopy; Linux: xclip)
  2. Opens the platform's compose / upload page in your browser (pre-filled with your caption on X, Reddit, and Facebook's sharer)
  3. Shows a one-line instruction ("Click + Create, then paste the file", etc.)

Text layers

A multiline caption selected, with the Text inspector open

Each caption is a clip on the timeline's Text track. Select one on the timeline (or with the arrows in the Inspector) and the Text tab edits it:

  • Style presets — Subtitle (white-on-black box), Caption (white with black outline, the social-standard look), Title (large centered), Watermark (small corner), Custom
  • Font (system fonts), size, color (8 presets + Custom… color picker), position (7 anchors, or drag it in the preview)
  • Bold / italic toggles, resolved to real font-variant files (arialbd.ttf, ariali.ttf, ...) with graceful fallback when a variant is missing
  • Outline, Shadow and Box toggles, compiled to drawtext borderw / shadowx / box and mirrored exactly in the preview
  • Multiline captions: the text box wraps, and embedded newlines export as real line breaks
  • Timing — type exact times, or drag the clip and its edges on the timeline
  • Record a motion path — keyframe a caption's position so it follows a moving subject; drag to set the path or Follow an object from here with OpenCV, then compiled to a drawtext time expression that preview and export share
  • Add, duplicate, and delete from the Inspector header
  • Fade (0.25s / 0.5s / 1s, for all captions) — symmetric fade-in/fade-out via a drawtext alpha= expression
  • Live PIL-rendered overlay on the preview canvas — font size, position, and box padding all match ffmpeg's export output

Quality presets

Preset FPS Max Width GIF Colors Video CRF
Low 10 480px 64 28
Medium 15 720px 128 23
High 24 1080px 256 18
Ultra 30 Native 256 15

When a hardware encoder is available, CRF maps to the right flag per encoder (-cq for NVENC, -global_quality for QSV, -q:v for VideoToolbox).


Keyboard shortcuts

Key Action
Space / K Play / Pause
J Move the playhead back 1s
L Move the playhead forward 1s
I Set the in point at the playhead
O Set the out point at the playhead
Q Save the selection as a range
Ctrl+Z Undo (captions, overlays, crop, selection, saved ranges)
Ctrl+Y / Ctrl+Shift+Z Redo
Ctrl+E Export with the current settings
Ctrl+O Open video file
Ctrl+V Paste — a video/GIF link opens Import from anywhere; a clipboard image becomes an overlay in Edit

Entry fields swallow shortcuts so typing into them doesn't scrub the video.


Installation

Option A — Windows (no Python required)

Get it from the Microsoft Store (signed, auto-updates, no security warning), or download VideoKidnapper.exe / the Setup installer from the latest release. If a prerequisite is missing, first-run setup explains the verified download and asks before installing it.

Option B — macOS (no Python required)

Download the .dmg for your Mac from the latest release — …-macos-arm64.dmg (Apple Silicon) or …-macos-x86_64.dmg (Intel) — and drag the app to Applications. FFmpeg is bundled.

Signed and notarized by Apple as of v1.8.2, so it opens with a normal double-click — no security prompt, no Settings trip.

If macOS says "Apple could not verify VideoKidnapper is free of malware", you have v1.8.1: press Done, then System Settings → Privacy & Security → Open Anyway, or just upgrade. If it says "VideoKidnapper is damaged and can't be opened", you have v1.8.0, whose DMGs shipped with a broken code signature that no workaround can clear — upgrade.

Option C — Linux AppImage (no Python required)

Download VideoKidnapper-x86_64.AppImage from the latest release:

chmod +x VideoKidnapper-x86_64.AppImage
./VideoKidnapper-x86_64.AppImage

FFmpeg is bundled — nothing else to install. Works on any glibc 2.35+ distro: Ubuntu 22.04+, Debian 12+, Fedora 36+, and immutable distros like Bazzite, SteamOS 3+, and Silverblue (where the AppImage is the recommended route since you can't layer packages). For an app-menu entry + icon, add it with Gear Lever.

On Ubuntu / Debian / Mint you can instead use the APT repository — one-time setup, then updates arrive through normal apt upgrade:

sudo install -d /etc/apt/keyrings
curl -fsSL https://aes256afro.github.io/apt/videokidnapper.asc | sudo tee /etc/apt/keyrings/videokidnapper.asc > /dev/null
echo "deb [signed-by=/etc/apt/keyrings/videokidnapper.asc] https://aes256afro.github.io/apt stable main" | sudo tee /etc/apt/sources.list.d/videokidnapper.list
sudo apt update && sudo apt install videokidnapper

(Or grab the .deb from the release page and sudo apt install ./videokidnapper_*.deb — same package, no repo setup, no auto-updates.)

PyPI also works if you prefer pip:

sudo apt install python3-pip python3-tk ffmpeg xclip
pip install "videokidnapper[all]"
videokidnapper
pip install videokidnapper            # core install
pip install "videokidnapper[dnd]"     # + drag-and-drop support
videokidnapper                        # launches the GUI
videokidnapper --help                 # CLI mode

You still need FFmpeg on PATH, or use Setup and components… in the app's settings menu to install a verified portable copy (Windows and macOS). On Linux use your package manager.

Option E — Clone and install (contributors / latest main)

1. Install Python 3.9 – 3.14

2. Clone and install

git clone https://github.com/AES256Afro/VideoKidnapper.git
cd VideoKidnapper
pip install -e .                      # editable install — picks up your edits
# or the old way:
pip install -r requirements.txt

3. FFmpeg

Three options — the Setup dialog handles all of them:

  • Auto-install (Windows and macOS): open Setup and components… from the settings menu → check FFmpeg → Install Selected. Pulls a verified build into assets/ffmpeg/bin/: gyan.dev essentials on Windows, the static build the Mac app bundles on macOS.
  • Manual portable: drop ffmpeg.exe and ffprobe.exe into assets/ffmpeg/bin/ yourself.
  • System install: winget install Gyan.FFmpeg (Windows) / brew install ffmpeg (macOS) / sudo apt install ffmpeg (Linux).

4. Run

python main.py           # GUI (works from a clone without installing)
python main.py --help    # CLI help
# or, after `pip install -e .`:
videokidnapper           # GUI
videokidnapper --help    # CLI help

CLI mode

python main.py --url "https://youtu.be/..." --start 10 --end 25 \
               --format GIF --quality High --speed 1.5 --aspect 9:16

All flags: --url, --file, --start, --end, --out, --format {MP4,GIF}, --quality {Low,Medium,High,Ultra}, --speed, --rotate {0,90,180,270}, --mute, --audio-only, --aspect {Source,1:1,9:16,16:9,4:5,3:4}, --hw {auto,off}.

Passing any flag (or --help) skips the GUI and runs headless.


Export naming

Exports are named after the video: the title a download reported, or the file name of a local video. Export ▸ File names picks the style and shows what the next file will be called:

Style Example
Video title (default) Cat Video Take 2.mp4
Video title + date Cat Video Take 2_20261006.mp4
Video title + date & time Cat Video Take 2_20261006_143022.mp4
VidKid + timestamp VidKid_trim_20261006_143022.mp4

Titles are cleaned up for the file system (path separators, Windows-illegal characters and device names like CON are removed; accents, CJK and emoji are kept), and a second export of the same clip gets _1, _2 rather than overwriting. Files go to the folder set in Export ▸ Save to.

When "Join everything into one video" is on, the intermediate per-range files are cleaned up and only the joined file is kept.


Tech stack

  • CustomTkinter — GUI framework (five themes: Studio, Graphite, Cream retro tech, Fallout, Retro)
  • Pillow — frame preview + live text-layer overlay
  • yt-dlp — multi-platform video downloading
  • FFmpeg — video/GIF encoding with drawtext overlays
  • mss — cross-platform screen capture
  • tkinterdnd2 (optional) — drag-and-drop on the preview canvas

Tests

pip install pytest
python -m pytest tests/ -v

500+ tests covering URL detection, project files and recovery paths, install-aware updates, verified prerequisite staging, platform share intents, ffmpeg filter construction, text styling, download retry classification and cookie resolution, settings migration, SRT parsing, size estimation, cache behavior, and drag-and-drop payload parsing.



Disclaimer & terms of use

VideoKidnapper is a personal utility for trimming and re-encoding video you have the right to use. By running this software you acknowledge:

  • Platform terms of service. Downloading from services like YouTube, Instagram, Twitter/X, Reddit, Bluesky, and Facebook may violate their terms of service. You are responsible for ensuring your use complies with each platform's current ToS, your local laws, and applicable copyright law (DMCA, Fair Use, and equivalents in your jurisdiction).
  • Copyright. Re-hosting, re-sharing, or monetizing content you don't own or have a license to use is your responsibility and not this project's.
  • Cookies. The "Cookies from browser" option reads authentication cookies from your installed browsers through yt-dlp. Never share a cookie file or export — it grants session-level access to your accounts.
  • No warranty. The software is provided "as is" without warranty of any kind; see the LICENSE file for the full terms.

The project's authors and contributors do not endorse or encourage violation of any platform's terms of service or any applicable law.


License

Licensed under the Apache License, Version 2.0 — see LICENSE for the full text.

Note on historic releases. The v1.0.0 tag and earlier were published under the GNU General Public License v3.0 and remain available under GPLv3. Apache-2.0 applies to v1.1.0 and every later commit on main.

VideoKidnapper depends on several third-party projects with their own licenses:

Project License
CustomTkinter MIT
Pillow MIT-CMU / HPND
yt-dlp Unlicense
mss MIT
tkinterdnd2 BSD-3-Clause (optional)
FFmpeg LGPLv2.1+ / GPLv2+ (build-dependent; essentials builds are GPL)

If you bundle FFmpeg with a redistribution of VideoKidnapper, note that the essentials and full gyan.dev builds ship under GPL — complying with GPL requires offering source on request. Alternatively, use an LGPL FFmpeg build for LGPL-only redistribution.

Metadata

Release files for videokidnapper 1.9.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for videokidnapper 1.9.1
File Size Uploaded
videokidnapper-1.9.1.tar.gz 1.0 MB Details

Built distribution (wheel)

Table of built distributions (wheels) for videokidnapper 1.9.1
File Interpreter ABI Platform
videokidnapper-1.9.1-py3-none-any.whl Python 3 none any Details

Total release size: 2.0 MB

Release files / videokidnapper-1.9.1.tar.gz

Download URL videokidnapper-1.9.1.tar.gz
Size 1.0 MB
Tags Source
SHA-256 checksum
How to use checksums
ad27906c98f2056f1cfb7a6c856f3961fb532ee4d2c79b5c530f5bf422f43823
BLAKE2b-256 checksum
How to use checksums
6eea0469e058b616225fbfdd9c3fd1b5a49711ccdbb10d19fe32438630b5db0b
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 8, 2026.

Transparency log

Release files / videokidnapper-1.9.1-py3-none-any.whl

Download URL videokidnapper-1.9.1-py3-none-any.whl
Size 962.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
dcef29a264b394f0d93c2f591c315dc3a989e5251f7a211dd05b29a943fd2153
BLAKE2b-256 checksum
How to use checksums
ea134a4765a5c382e39c5a1761b863fa9bda72a651b95114bd06020d40647fea
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 8, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

1.9.1 This release

2 release files

1.9.0

2 release files

1.8.4

2 release files

1.8.3

2 release files

1.8.2

2 release files

1.8.1

2 release files

1.8.0

2 release files

1.7.4

2 release files

1.7.3

2 release files

1.7.2

2 release files

1.7.1

2 release files

1.7.0

2 release files

1.6.0

2 release files

1.5.1

2 release files

1.5.0

2 release files

1.4.1

2 release files

1.4.0

2 release files

1.3.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page