videobeaux manual

videobeaux manual

How to use videobeaux, then every program mode and its settings.

Start Here

Welcome

videobeaux is a video playground. Drag program modes onto a canvas, connect them, press Run.

There are 168 program modes, grouped by what they do: glitch, trails, time, colour, tracking, layout, editing, captions and more. You can use one on its own or chain many, each feeding the next. Everything runs on your computer, offline, and your videos stay there.

New here? Do Install then Getting started. It takes about two minutes.

Start Here

Install

videobeaux has everything built in. Nothing else to install.

Mac

You need a Mac with Apple silicon (M1 or later).

  1. Unzip the download and drag videobeaux into your Applications folder.
  2. Open it once from Applications. macOS says it can't be opened. Click Done.
  3. Open System Settings → Privacy & Security, scroll to the message about videobeaux and click Open Anyway.

This happens because the app isn't signed with an Apple certificate yet. It's safe to open, and it only asks once.

Windows

You need Windows 10 or 11 (64-bit) and about 4 GB of free space.

  1. Run the videobeaux Setup file you were sent.
  2. Windows says Windows protected your PC. Click More info, then Run anyway.
  3. Follow the installer. The default folder is fine.

First launch

A Setup screen checks the built-in video tools. Speech, voice and background-removal features download their models the first time you use them. Some are large.

Start Here

Getting started

  1. Pick your video in the blue Input node (Browse…).
  2. Add programs: drag one from the list on the left onto the canvas, or double-click its name.
  3. Connect them: click the dot on the right edge of one box, then click the dot on the left edge of the next. A line follows your cursor in between. Chain it Input → programs → Output.
  4. In the orange Output node choose where to save (Save As…) and a format.
  5. Press ▶ Run Pipeline. Progress shows in the Signal Out bar at the bottom.

Quick first try: add a short video, search the program list for bad, drag Bad Predator onto the canvas, connect Input → Bad Predator → Output, and press Run.

Start Here

Handy to know

  • Branching & extra videos — one output can feed several programs. Programs that need a second video (Layer Blend, Stack 2×…) show an extra dot, or drag + Input video onto the canvas for another source.
  • Numbers — drag the slider or type a value (the box can be emptied and retyped). The faint strip under a slider is a good starting range; the black tick is the default; ↺ resets. Some programs have Presets chips that fill several values at once. A slider marked curved (hover) is bent so its middle gives about half the visible effect; the box always shows the real value.
  • ✨ Tidy (⌘⇧L) arranges a messy canvas in connection order, or only the selected programs. It runs left → right or top → bottom, whichever suits the shape of the window (⌥⌘⇧L always goes top → bottom); ⌘Z undoes it. The toolbar shrinks with the window — ◂ hides it, ☰ Tools holds everything when it is narrow.
  • Did it work? — with the console folded, a chip shows ✓ Done, ✗ Failed at step n or ■ Cancelled; click it for the log. A notification appears if the window is in the background (🎨 Appearance → Notify when done).
  • Select several — ⌘/Ctrl-click or Shift-drag a box around them. ⌘G groups them so they move together; the caret on a group folds it into one card.
  • Copy & paste — ⌘C / ⌘V copy programs and the connections between them; ⌘D duplicates. ⌘Z undoes (also after Clear).
  • Clear empties the canvas after asking you to confirm. Save/Load Preset keeps a whole setup.
  • Fold a program — ▾ (or double-click its header) hides its settings; ▤ / ▦ in the bottom-left controls fold or unfold all. Folded state is saved in presets.
  • Look & lines — 🎨 Appearance changes the Look (Classic is the original), the colour of connection lines, and puts an image behind the canvas.
  • Visual editors — a ✎ button next to a field opens a purpose-built editor, such as Lagkage → Layout JSON (drop in images).
Start Here

Keyboard shortcuts

On Windows, use Ctrl in place of ⌘.

KeysWhat it does
Delete / BackspaceDelete the selected programs or connection
⌘ASelect everything
⌘C ⌘V ⌘DCopy, paste, duplicate
⌘G / ⇧⌘GGroup / ungroup
⌘Z / ⇧⌘ZUndo / redo
⇧⌘LTidy: auto-arrange the canvas
⌥⇧⌘LTidy top → bottom
EscCancel a connection in progress, or clear the selection
⌘BShow / hide the program list
⌘JShow / hide the console
F1 or ?Open this help
Start Here

When something goes wrong

  • Read the Signal Out console — the last red line usually says what happened.
  • Nothing runs, or a red “not ready” message: open ⚙ Setup and press Repair.
  • Phone or HDR clips misbehaving: run Convert first (to mp4), then use that file.
  • AI features (speech, narration, background removal) need a one-time download in Setup → Local AI features; afterwards everything works offline.
  • A field says “max …”: that value is out of range and was set to the limit.
  • Found a bug? Click the ? icon in videobeaux to make a bug report, then send it on the feedback page.
Start Here

Licenses & source

  • Videobeaux runs ffmpeg as a separate program. The bundled build is a GPL build (libx264 / libx265) that is never built with the non-redistributable --enable-nonfree option.
  • The exact source code of that ffmpeg and every library in it is published alongside each release of the build; the licence summary ships in the app's LICENSES folder.
  • You can use your own ffmpeg instead (Setup). AI models you download (speech, narration, background removal) keep their own licences.
Glitch & Corruption

Bad Animation

introduces uneven timing to simulate poorly animated sequences.

Applies a “bad animation” effect to the input video — producing jittery, uneven, low-frame-feeling, or intentionally crude animated motion. This effect simulates broken or poorly interpolated animation, often seen in retro cartoons, limited-animation TV shows, or stylized glitch art.

Good for: Music videos, experimental film, and stylized animation · Retro cartoon emulation or low-budget limited-animation effects · Distortion-based visual art, glitchworks, or datamosh-adjacent styles

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Pair with Frame Interpolate (ironically) to contrast smooth vs broken motion.

Glitch & Corruption

Bad Contrast

Harsh blend-mode contrast corruption — hard-mix, vivid-light and exclusion blends over a softened copy.

Applies a “bad contrast” visual effect to the video, intentionally degrading tonal balance, highlight rolloff, shadow separation, and midtone neutrality. The result simulates low-quality analog transfers, misconfigured broadcast equipment, degraded VHS dubs, or broken grading pipelines.

Good for: Broken-TV or analog-digital hybrid aesthetics · Glitch art, experimental cinema, and abstract video installations · Distressing otherwise clean digital footage

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Apply at the end of a chain to maximize visible distortion.

Glitch & Corruption

Digital Boss

Hue/saturation shift plus extreme frame-difference amplification — a busted, blown-out digital look.

Applies a “busted Game Boy–style digital boss” effect to the video — producing harsh bitmap artifacts, limited 8-bit tonal mapping, flicker, false edges, and unstable block transitions reminiscent of corrupted handheld-console graphics. The effect aims to evoke a glitched, retro, boss-fight visual language.

Good for: Retro boss-fight sequences or video-game-themed edits · Concert visuals or projection loops with digital-grunge aesthetics · Glitch-art animation, collage, and experimental video

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Convert Dims to produce Game-Boy-like square aspect ratios.

Glitch & Corruption

XRGB

Extreme RGB adjustment.

Applies an extreme RGB color treatment that aggressively remixes channel intensity, contrast, and chromatic balance. This effect pushes the video into hyper-saturated, unnatural, or neon-driven territory — producing explosive color distortions ideal for glitch, psychedelia, or high-impact stylization.

Good for: Psychedelic music videos · Glitch-art edits requiring aggressive color disruption · Abstract collage or RGB-driven montage

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Slight Smear for neon-blooming distortion.

Glitch & Corruption

Crossmosh

Cross-mosh (real datamosh): keep decoder state A→B; optional smear trail.

A controlled datamoshing engine that manipulates motion vectors, frame ordering, and predictive compression structures to create stylized glitch-drift distortions. Unlike chaotic or accidental datamoshing, Crossmosh provides intentional control over how macroblocks bleed, drift, smear, or desynchronize across frames.

Good for: Music videos requiring controlled glitch cascades · VJ loops and live visual sets that leverage drifting corruption · Collage work blending two unrelated videos via motion-vector infection

Settings

SettingWhat it does
B Clip (Second Video)The second clip that gets datamoshed into Connect a second video to its extra dot, or add one with + Input video.
Codeclibxvid gives better results but needs an ffmpeg build with it — most Homebrew installs don't have it
Default: mpeg4 · Choices: libxvid, mpeg4
ModeDefault: proto · Choices: proto, smear
Q ScaleQuality scale — lower is better
Default: 3 · Range 1 to 31
GOP SizeGOP interval (keyframe distance)
Default: 9999 · Range 1 to …
Smear Frames[smear] tmix frame count
Default: 9 · Range 1 to …
Smear Decay[smear] lagfun decay 0–1
Default: 0.9 · Range 0 to 1

Tip: Longer GOPs (100–300+) produce dramatic streak-based artifacts across many frames.

Glitch & Corruption

Pixel Sort

Glitch-art pixel sorting — bright pixels smear into sorted streaks.

Settings

SettingWhat it does
Threshold (0-255, empty = auto)Brightness above (or below) which pixels get sorted. Leave empty for AUTO: the threshold follows each frame's own brightness, so the picked share stays about the same from clip to clip. Lower = more pixels sorted.
Range 0 to 255
Sort pixels that arebrighter = bright areas smear (default). darker = shadows smear instead.
Default: brighter · Choices: brighter, darker
Sort byWhat orders the pixels inside a streak. brightness is the classic look; hue, saturation or a colour channel give banded colour streaks.
Default: brightness · Choices: brightness, hue, saturation, red, green, blue
Orderascending = low values first (streaks run up to bright). descending = the reverse.
Default: ascending · Choices: ascending, descending
Vertical streaksSort down the columns instead of across the rows.
Shortest streak (px)Streaks shorter than this stay as they are. 1 = sort everything picked; 8–30 keeps only long streaks.
Default: 1 · Range 1 to 400 · Good start: 1–60
Mix with original1 = fully sorted. Lower blends the sorted picture back over the original.
Default: 1 · Range 0 to 1 · Good start: 0–1
Downscale for speed (px, 0 = off)Sort at this width to go faster on 4K footage — the result is scaled back up to the original size, so the output is never smaller than the input. 0 = no downscaling.
Default: 0 · Range 0 to 4096 · Good start: 0–1920
Glitch & Corruption

VHS Tracking Error

Classic worn-VHS-tape look: horizontal tracking-jitter wobble plus chroma bleed/smear, like a tape that's been rewound one too many times.

Settings

SettingWhat it does
IntensityOverall strength multiplier for the chroma bleed, grain, and jitter. Default: 1.0.
Default: 1 · Range 0 to 1
SpeedWobble frequency — higher jitters faster. Default: 6.0.
Default: 6 · Range 0 to 30
Glitch & Corruption

Chromatic Pulse

Animated chromatic aberration: red and blue channels pulse apart from green and back again over time, instead of sitting at a fixed static offset.

Settings

SettingWhat it does
AmountMaximum pixel shift at the peak of the pulse. Default: 6.0.
Default: 6 · Range 0 to 30
SpeedPulses per second, roughly. Default: 2.0.
Default: 2 · Range 0 to 10
Glitch & Corruption

Deep Fry

The 'deep-fried meme' look: blown-out saturation and contrast, oversharpened, baked-in compression artifacts.

Settings

SettingWhat it does
IntensityOverall strength multiplier. Default: 1.0.
Default: 1 · Range 0 to 1
Glitch & Corruption

Twociz

Apply filter from the perspective of a zombie on TC-1 hallucinogens.

Applies a distorted, delirious visual effect meant to simulate the hallucinated perspective of a zombie under the fictional TC-1 compound. The aesthetic is disoriented, chemically corrupted, semi-conscious, and grotesquely vivid — combining abnormal color shifts, lurching motion, and sensory confusion.

Good for: Horror or undead-themed sequences · Music videos requiring surreal, lurching, toxic hallucinations · Psychological visualizations of decay, delirium, intoxication, or inner collapse

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Pair with Septic for medical-shock surrealism.

Glitch & Corruption

Twociz Pro

Apply filter from the perspective of a zombie on TC-1 hallucinogens.

Settings

SettingWhat it does
Radius (px)Neighborhood size for amplify. Small = sharp/edgy; large = broad smeary hallucination. Try 2–8.
Default: 4 · Range 1 to 32 · Good start: 2–8
Amplify strengthAmplify strength. 1.0 is mild; higher values get more intense/crunchy. Try 1.2–3.0.
Default: 2 · Range 0.5 to 10 · Good start: 1.2–3
Key edge softnessChromakey edge softness. 0 = hard edge; higher = feathered edge. Try 0.0–0.20.
Default: 0.1 · Range 0 to 1 · Good start: 0–0.2
Blue toleranceChromakey tolerance for blue. Higher removes more blues (and can eat nearby colors). Try 0.10–0.35.
Default: 0.2 · Range 0.01 to 1 · Good start: 0.1–0.35

Presets: Gentle Medium Wild

Glitch & Corruption

Splitting

Rearranges pixels vertically and inversely. Looks kind of like a damaged tape.

A simple segmentation utility that divides a video into evenly timed chunks or scene-based fragments for modular editing. Unlike Splitting Pro, which slices spatial regions of the frame, Splitting operates purely in the time domain, cutting the clip into temporal sections that can be reused, rearranged, or processed individually.

Good for: Procedural editing pipelines · Video collage systems that reassemble timelines randomly · Creating modular “building blocks” for generative video art

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Apply Convert Dims first if you want uniformity across all segments.

Glitch & Corruption

Splitting Pro

Rearranges pixels vertically and inversely. Looks kind of like a damaged tape.

A precise segmentation tool that slices a source video into reusable chunks based on positional slicing rules. Splitting Pro allows creators to extract vertical or horizontal bands of the frame, enabling compositing, collage workflows, displacement patterns, and region-based editing.

Good for: Split-screen collage layouts · Creating stripe-based glitch effects · Feeding sliced layers into Lagkage for recombination

Settings

SettingWhat it does
WidthSize in pixels of the shuffled slices. Default: 157.
Default: 157
PositionShuffle direction: vertical slices, horizontal slices, or square blocks. Default: vertical.
Default: vertical · Choices: vertical, horizontal, block

Tip: Pair with Blur Pix for soft-edge slicing.

Glitch & Corruption

Slight Smear

Slightly smearing RGB color space.

Applies a subtle RGB smear effect, gently shifting and bleeding color channels to create a soft distortion. The look resembles mild chromatic drift, analog smudge, or a barely-there color echo — perfect for adding texture without overwhelming the original image.

Good for: Music videos needing gentle stylistic color bloom · Dreamy ambience sequences and soft-focus aesthetics · Collage art where slight distortion adds cohesion

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Gamma Fix beforehand for more stable channel balancing.

Glitch & Corruption

Blur Pix

Applies a double-layered blur and pixelation combined effect.

Applies a pixelated blur effect to the input video, producing a blocky, low-resolution softening reminiscent of censored footage, retro 8-bit graphics, or lo-fi compression artifacts. This effect blends coarse pixelation with smooth blur, creating a stylized, abstracted visual aesthetic.

Good for: Music videos needing stylized abstraction · Identity-obscuring or anonymization with artistic flair · Retro-inspired or video-game-like sequences

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Pair with Hash Fingerprint or Bad Contrast to stack analog/digital degradation.

Glitch & Corruption

Warp

Geometric warps: swirl, bulge, pinch, ripple, kaleidoscope and mirror modes, optionally animated. Fast (precomputed remap).

Settings

SettingWhat it does
WarpWarp type. Default: swirl.
Default: swirl · Choices: swirl, bulge, pinch, ripple, kaleidoscope, mirror_h, mirror_v, mirror_quad
AmountEffect strength. Default: 1.
Default: 1 · Range 0 to 1
RadiusAffected radius as a fraction of the frame (swirl/bulge/pinch). Default: 0.7.
Default: 0.7 · Range 0.05 to 1.5
Center XCenter X, 0-1. Default: 0.5.
Default: 0.5 · Range 0 to 1
Center YCenter Y, 0-1. Default: 0.5.
Default: 0.5 · Range 0 to 1
SegmentsKaleidoscope segments. Default: 6.
Default: 6 · Range 2 to 32
SpeedAnimation speed (0 = static). Swirl oscillates, ripple travels, kaleidoscope rotates.
Default: 0 · Range -5 to 5
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Glitch & Corruption

Flow Warp

Optical-flow smear/push — pixels drag along motion like a datamosh melt — or a color flow view.

Settings

SettingWhat it does
Modesmear = pixels keep dragging along motion (datamosh-like), push = current frame warped by its motion, flow_view = motion shown as color.
Default: smear · Choices: smear, push, flow_view
StrengthHow far pixels are pushed along the motion. Default: 3.
Default: 3 · Range 0 to 15
PersistenceSmear: how long smeared pixels persist, 0-0.99. Default: 0.9.
Default: 0.9 · Range 0 to 1
Flow WidthWidth the motion is analyzed at (smaller = faster). Default: 320.
Default: 320 · Range 96 to 1280
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Glitch & Corruption

Glitch Tear

RGB channel split plus horizontal tears and scanline dimming, re-rolled every frame.

Settings

SettingWhat it does
IntensityDefault: 0.5 · Range 0 to 1
Tears per frameDefault: 4 · Range 0 to 16
RGB shift (px)Default: 9 · Range 0 to 40
No scanlines
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Glitch & Corruption

Weak Signal

Edge-of-reception transmission: skew, colour misregistration, noise, streaks and bursts of static.

Settings

SettingWhat it does
NoiseDefault: 14 · Range 0 to 60
StreaksDefault: 5 · Range 0 to 30
Static burstsChance per frame.
Default: 0.35 · Range 0 to 1
SkewDefault: 0.035 · Range 0 to 0.2
SpeckleDefault: 0.02 · Range 0 to 0.2
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Glitch & Corruption

VHS Camcorder

Home-video camcorder look: soft chroma, tape noise, a rolling tracking band and the on-screen PLAY / date stamp.

Settings

SettingWhat it does
On-screen labelDefault: PLAY
Hide on-screen display
Hide date/time
Tape noiseDefault: 7 · Range 0 to 30
No tracking band
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Glitch & Corruption

CRT Monitor

Old TV tube: curved glass, scanlines, RGB shadow mask, glow and a dark vignette.

Settings

SettingWhat it does
CurvatureDefault: 0.1 · Range 0 to 0.4
ScanlinesDefault: 0.45 · Range 0 to 1
Shadow maskDefault: 0.28 · Range 0 to 0.8
GlowDefault: 0.3 · Range 0 to 1
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Trails & Echoes

Ghostee

Amplifies visual features or distortions for a spiritual-esque experience.

Applies a soft, subtle ghosting effect to the video by blending frames over time. This creates a trailing, dreamy afterimage that lightly smears motion and adds a supernatural or ethereal visual mood.

Good for: Dreamlike music videos or ambient visualizers · Supernatural, memory-like, or nostalgic sequences · Layered VJ and projection-mapped installations

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Double Cup for syrupy dream visuals.

Trails & Echoes

LSD Feedback

Wicked trippy bro.

Applies a psychedelic recursive feedback effect inspired by analog video feedback loops and LSD-style visual hallucination. The effect blends delayed frames into the live frame stream, producing spiraling trails, melting motion, neon echoes, and evolving kaleidoscopic distortions.

Good for: Psychedelic music videos, rave visuals, and festival projections · VJ loops with evolving feedback tunnels · Experimental films exploring self-referential distortion

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Frame Delay Pro 1 or Frame Delay Pro 2 to produce layered temporal echoes.

Trails & Echoes

LSD Feedback Pro

Wicked trippy bro.

Settings

SettingWhat it does
Frames mixedNumber of frames to mix across time. Higher values create longer motion trails and heavier ghosting. Must be >= 8. Try 8–16.
Default: 8 · Range 8 to 32 · Good start: 8–16
Trails & Echoes

Double Cup

Apply the effect of purple drank. Yum.

Applies a syrupy, slowed, woozy audiovisual effect reminiscent of “purple drank” aesthetics. The look emphasizes heavy color bleeding, slowed-feeling motion, chromatic haze, and a dreamlike viscosity often associated with chopped-and-screwed culture or psychedelic video distortion.

Good for: Vaporwave, cloud rap, chopped-and-screwed visuals · Psychedelic edits and dream-sequence overlays · Live VJ loops and projection-mapped environments

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Blur Pix or Bad Contrast for more degraded, syrup-soaked textures.

Trails & Echoes

Mirror Delay

Apply a frame delay plus a mirrored effect.

Applies a mirrored frame-delay effect, blending temporal echoes with reflective symmetry. The result combines soft motion trailing with a mirrored duplication of the frame, creating hypnotic, kaleidoscopic, or hall-of-mirrors–style distortion.

Good for: Music videos with mirrored choreography or symmetrical staging · Psychedelic pieces needing a dreamlike reflective effect · Ambient loops and visualizations requiring geometric repetition

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Pair with Convert Dims to square or portrait layouts for perfect symmetry.

Trails & Echoes

Frame Delay Pro 1

Makes the video seem like a dream when you are sick with a bug.

Applies an advanced frame-delay effect using a customizable number of delayed frames and weighted blending. This produces echoing motion trails, temporal smearing, ghosted motion artifacts, and stylized feedback-like delay effects.

Good for: Music video trails and rhythmic pulse effects · Rotoscoped animations requiring smeared or echoed movement · Live VJ feedback systems and performance visuals

Settings

SettingWhat it does
Frame QuantityNumber of past frames to retain to control length of visual memory trail.
Frame WeightsValue between 0.0 and 1.0. How much influence each retained frame has when blended into the output.

Tip: Normalize weights so the sum stays below ~2.0 if you want clean blending without clipping.

Trails & Echoes

Frame Delay Pro 2

Makes the video seem like a dream when you are sick with a bug.

Applies an advanced analytical frame-delay effect that shifts image planes (luma/chroma) independently while applying temporal decay. This creates ghost trails, chromatic drifts, smeared movement, and layered motion persistence that evolves across time.

Good for: Music videos needing rhythmic or drifting trail effects · Abstract animations or temporal collage art · Footage where separating luma and chroma drift enhances surrealism

Settings

SettingWhat it does
DecayValue between 0.0 and 1.0. Controls how quickly the trail fades.
PlanesValue is a bitmask (0–7) for YUV planes. 0 = no effect 1 = Y (luma/brightness) 2 = U (blue chroma) 3 = Y + U Brightness + blue projection 4 = V (red chroma) 5 = Y + V Brightness + red projection 6 - U + V Full color, but not brightness 7 - Y + U + V Affects everything (full color trail)

Tip: Combine with Frame Delay Pro 1 for multi-layer delay stacks.

Trails & Echoes

Fever

Makes the video seem like a dream when you are sick with a bug.

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1
Trails & Echoes

T-1000

Temporal median smoothing with RGB offsets — liquid-mercury shimmer on motion.

Applies a liquid-metal distortion effect inspired by the T-1000 from Terminator 2. This effect simulates molten chromium surfaces, flowing reflections, metallic warping, and fluid reshaping — producing a highly stylized morphing look that feels synthetic, slippery, and impossibly smooth.

Good for: Sci-fi transformations or robotic metamorphosis sequences · Music videos seeking surreal chrome or mercury aesthetics · Abstract collage or experimental edits involving fluid distortion

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1
Working detail (px)The effect is computed at this picture height, then scaled back up. Lower = much faster (540 runs ≈1.5× real time on HD and 4K; full size is ≈0.7× on HD and ≈0.2× on 4K). Higher = sharper still parts.
Default: 540 · Range 0 to 2160 · Good start: 360–1080

Presets: Fast (360) Balanced (540) Sharper (720) Full size

Tip: Combine with Steel Wash for a colder, industrial chrome palette.

Trails & Echoes

Smudge

Introduces motion-like smudges for a soft, blended output.

Applies a gentle smudge effect to the video, softening edges and pulling colors slightly across motion paths. The result resembles a faint manual smear, as if the image were lightly dragged or brushed while still wet.

Good for: Dream sequences, surreal atmospheres, or emotional montage · Soft, analog-inspired grading setups · Background layers in collage or Lagkage layouts

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Slight Smear for layered organic softness.

Trails & Echoes

Repainting

Akin to repainting the same image while smudged with alcohol.

Applies an artistic “repainted” effect that makes each frame appear as if it has been smeared, redrawn, or blurred by an alcohol-washed brush. The result resembles wet ink, soft smudging, streaked pigment, or repeated overpainting — giving the footage a tactile, analog, hand-altered feel.

Good for: Abstract or painterly music videos · Alcohol-ink inspired motion design · Narrative moments needing dreamlike or smeared transitions

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Double Cup for a liquified wash; combine with Blur Pix for soft surface blending.

Trails & Echoes

Long Exposure

A slowly fading average of past frames — moving things smear into ghostly trails.

Settings

SettingWhat it does
PersistenceHigher = longer trails.
Default: 0.85 · Range 0 to 0.99
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Trails & Echoes

RGB Time Split

Red channel is now, green and blue lag behind — motion leaves rainbow fringes.

Settings

SettingWhat it does
Green delay (frames)Default: 9 · Range 0 to 60
Blue delay (frames)Default: 18 · Range 0 to 60
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Trails & Echoes

Feedback Loop

Video feedback: the output is fed back into the mix, zoomed, rotated, shifted and hue-shifted a little each frame, spiraling the picture into itself like pointing a camera at its own monitor.

Settings

SettingWhat it does
Feedback amountHow much of the previous output returns. Near 1 = long, endless trails.
Default: 0.7 · Range 0 to 1
Zoom (% per frame)Positive zooms the echo outward into a tunnel; negative shrinks it inward.
Default: 2 · Range -20 to 20
Rotate (° per frame)Rotation per frame in degrees. Default: 1.
Default: 1 · Range -30 to 30
Shift X (px)Horizontal shift per frame in pixels. Default: 0.
Default: 0 · Range -100 to 100
Shift Y (px)Vertical shift per frame in pixels. Default: 0.
Default: 0 · Range -100 to 100
Hue shift (° per frame)Hue shift per frame in degrees. Default: 4.
Default: 4 · Range -45 to 45
CombineHow the live picture joins the feedback.
Default: screen · Choices: screen, lighten, add, mix, over (keyed by brightness)
Key thresholdFor 'over (keyed)': only live pixels brighter than this are drawn over the feedback.
Default: 90 · Range 0 to 255

Presets: Tunnel Spiral Melt

Trails & Echoes

Key Feedback

Video feedback with two keys — a luma/color Insert key picks what enters the loop, a Feedback key picks what survives each pass — plus zoom, rotate, shift and hue shift.

Settings

SettingWhat it does
Insert keyWhich part of the live picture is let into the loop.
Default: luma · Choices: none, luma, color, luma AND color, luma OR color
Insert: luma fromBrightness band that enters the loop (0 = black, 1 = white).
Default: 0.5 · Range 0 to 1 · Good start: 0.3–0.6
Insert: luma toInsert luma band end, 0 to 1. Default: 1.
Default: 1 · Range 0 to 1 · Good start: 0.8–1
Insert: edge softnessInsert edge softness. Default: 0.1.
Default: 0.1 · Range 0 to 0.5 · Good start: 0.02–0.2
Insert: colorInsert key color. Default: #00FF00.
Default: #00FF00
Insert: color rangeHow far from the picked color still counts (hue/chroma only, brightness ignored).
Default: 0.3 · Range 0.02 to 1 · Good start: 0.15–0.5
Insert: invertLet in everything except the key.
Feedback keyWhich part of the recirculating picture survives each pass.
Default: none · Choices: none, luma, color, luma AND color, luma OR color
Feedback: luma fromFeedback luma band start. Default: 0.15.
Default: 0.15 · Range 0 to 1 · Good start: 0.1–0.4
Feedback: luma toFeedback luma band end. Default: 1.
Default: 1 · Range 0 to 1 · Good start: 0.7–1
Feedback: edge softnessFeedback edge softness. Default: 0.1.
Default: 0.1 · Range 0 to 0.5 · Good start: 0.02–0.2
Feedback: colorFeedback key color. Default: #FF00FF.
Default: #FF00FF
Feedback: color rangeFeedback color range. Default: 0.3.
Default: 0.3 · Range 0.02 to 1 · Good start: 0.15–0.5
Feedback: invertKeep everything except the key.
Feedback amountHow much of the previous output returns each pass. Near 1 = long trails.
Default: 0.9 · Range 0 to 0.99 · Good start: 0.6–0.95
Zoom (% per frame)Zoom per frame in percent. Default: 1.5.
Default: 1.5 · Range -20 to 20 · Good start: -3–4
Rotate (° per frame)Rotation per frame in degrees. Default: 1.
Default: 1 · Range -30 to 30 · Good start: -3–3
Shift X (px)Horizontal shift per frame, pixels. Default: 0.
Default: 0 · Range -100 to 100
Shift Y (px)Vertical shift per frame, pixels. Default: 0.
Default: 0 · Range -100 to 100
Hue shift (° per frame)Hue shift per frame in degrees. Default: 4.
Default: 4 · Range -45 to 45 · Good start: -8–8
Combineover = keyed pixels replace the feedback; the others blend them in.
Default: over · Choices: over, screen, lighten, add
ViewShow a matte instead of the result to tune the keys (white = in).
Default: output · Choices: output, insert key, feedback key

Presets: Bright trails Color echo Dark tunnel

Time & Motion

Speed

Changes the speed of the video file, from 0.05× (ultra slow-motion) to 100×. The audio follows the same speed without changing its pitch (or can be muted).

Good for: Music video edits synced to rhythmic patterns · Timelapse creation · Slow-motion mood sequences

Settings

SettingWhat it does
Speed (×)1 = normal · below 1 slows down (0.1 = ten times slower) · above 1 speeds up. The slider is logarithmic so slow and fast get equal room; type any value from 0.05 to 100.
Default: 1 · Range 0.05 to 100 · Good start: 0.25–4
AudioStretch keeps the pitch and follows the new speed (a drone at 0.1×) · Mute drops the sound.
Default: Stretch (keep pitch) · Choices: Stretch (keep pitch), Mute
Smooth slow-motionInvents in-between frames instead of repeating frames when slowing down. Much nicer at 0.1–0.5×, much slower to render.

Presets: Ultra slow (0.1×) Slow-mo (0.25×) Half speed Double Time-lapse (8×)

Tip: For the cleanest slow motion, run Frame Interpolate before Speed to create missing frames.

Time & Motion

Reverse

Reverses the video.oediv eht sesreveR.

Reverses the playback of the input video, producing a backward-motion effect. This simple but powerful manipulation can completely change the emotional or aesthetic tone of footage, turning ordinary actions into surreal rewinds, resets, or temporal loops.

Good for: Rewind effects in music videos or narrative sequences · Loop creation by combining forward + reversed versions of the same clip · Glitch-art assemblies where time feels elastic or broken

Settings

SettingWhat it does
AudioReversed = the sound plays backwards with the picture (default) · Forward = keep the sound as it was · Mute = no sound.
Default: Reversed · Choices: Reversed, Forward, Mute

Tip: Pair with Looper Pro to generate seamless mirrored loops.

Time & Motion

Boomerang

Forward-then-reverse ping-pong loop — the clip plays normally, then plays itself backwards, back to the start.

Settings

SettingWhat it does
Back-and-forthsHow many forward-then-backward cycles to play (1 = a single ping-pong).
Default: 1 · Range 1 to 6
Time & Motion

Time Ramp

Variable speed ramp within one clip (e.g. slow-mo into a speed-up), unlike the flat Speed effect.

Settings

SettingWhat it does
Start FactorSpeed factor at the start of the clip. Default: 0.5 (half speed).
Default: 0.5 · Range 0.1 to 8
End FactorSpeed factor at the end of the clip. Default: 2.0 (double speed).
Default: 2 · Range 0.1 to 8
StepsNumber of discrete speed steps making up the ramp. More = smoother. Default: 6.
Default: 6 · Range 2 to 30
Time & Motion

Freeze Punch

Freezes on detected audio peaks then resumes — a punchy freeze-frame emphasis edit.

Settings

SettingWhat it does
CountNumber of freeze points to insert. Default: 4.
Default: 4 · Range 1 to 20
Freeze DurationHow long each freeze holds, in seconds. Default: 0.4.
Default: 0.4 · Range 0.05 to 5
Min GapMinimum seconds between freeze points. Default: 3.0.
Default: 3 · Range 0 to 30
Scan WindowWindow size for the loudness scan, in seconds. Default: 0.5.
Default: 0.5 · Range 0.1 to 5
Time & Motion

Strobe Cut

Periodic flash/strobe effect — brightness spikes to near-white for a brief window at a configurable interval.

Settings

SettingWhat it does
PeriodSeconds between flashes. Default: 0.5.
Default: 0.5 · Range 0.05 to 10
Flash DurationHow long each flash lasts, in seconds. Default: 0.05.
Default: 0.05 · Range 0.01 to 2
Time & Motion

Stutter Pro

Imagine watching a video where random frames are played instead of a smooth progression.

Applies a customizable frame-stutter effect where playback cadence is intentionally disrupted according to a user-defined stutter intensity.

Good for: Music videos requiring beat-sync visual pulses · Experimental collage sequences with mechanical or robotic movement · Hyper-stylized glitch breakdowns

Settings

SettingWhat it does
Stutter window (frames)Replaces the current video frame with a randomly selected one from the most recent N frames.The larger the value, the larger the variation.
Default: 5 · Range 2 to 60 · Good start: 3–12

Tip: Slight blur (Blur Pix) can soften the harshness for dreamy styles.

Time & Motion

Nostalgic Stutter

Apply frame stutter akin to a corrupted file.

Applies a retro, analog-inspired frame-stutter effect reminiscent of corrupted home movies, aging videotape playback, or dropped-frame digital artifacts. Nostalgic Stutter intentionally disrupts cadence to evoke glitchy playback, uneven frame advancement, or mechanical jitter from vintage camcorders.

Good for: Vintage-style edits or memory sequences · Music videos needing lo-fi temporal distortion · Glitch-montage transitions and staccato cuts

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Bad Contrast or Double Cup for stylized aging or dreamy deterioration.

Time & Motion

Overexposed Stutter

Apply a frame stutter and exposing the video like the file is corrupted.

Applies a harsh overexposure flicker combined with stuttering frame cadence, producing an aggressive “corrupted camera” aesthetic. The effect mimics the look of damaged digital sensors, broken shutter timing, blown-out highlights, and unstable frame playback.

Good for: High-energy glitch sequences in music videos · Horror, sci-fi, or techno-thriller aesthetics involving corrupted vision · Fast-cut montage transitions

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Bad Contrast to push black levels down while highlights blow out.

Time & Motion

Looper Pro

Apply video looper effect base on frame size & start frame.

Creates a professional looping effect by selecting a region of frames and replaying them in a seamless or stylized loop. Looper Pro allows the user to choose which part of a video becomes a loop, enabling rhythmic repetition, GIF-like behavior, or precisely timed visual cycles.

Good for: Ambient loops for projection mapping or VJ sets · Social-media loops where infinitely repeating visuals are desirable · Narrative interludes or transitions requiring rhythmic repetition

Settings

SettingWhat it does
Segment length (frames)The number of frames in the segment to loop.
Default: 12 · Range 1 to 300 · Good start: 6–48
RepeatsNumber of additional times to repeat the segment. 2 means the segment appears 3 times total (original + 2 repeats).
Default: 2 · Range 1 to 20 · Good start: 1–6
Start frameThe starting frame number in the video where the segment begins.
Default: 0 · Range 0 to 100000 · Good start: 0–2000

Tip: Pre-process with Convert Dims for perfectly square or vertical loop assets.

Time & Motion

Frame Interpolate

Paints in-between frames: smoother motion (30→60 fps) or smooth slow motion. Slow on long clips.

Generates smooth slow-motion or higher-FPS video by creating intermediate frames using motion-compensated interpolation.

Good for: Slow-motion cinematic sequences · High-FPS playback for sports, action, and dance footage · Animation smoothing for stylized or experimental outputs

Settings

SettingWhat it does
Frame rate multiplier2 = twice as many frames per second (30 → 60: smoother motion). 1 = keep the frame rate (use with Slow motion).
Default: 2 · Range 1 to 8
Slow motion1 = normal speed. 2 = half speed, 4 = quarter speed — the extra frames are painted in so it stays smooth.
Default: 1 · Range 1 to 16
Target FPS (optional)Set an exact output frame rate (e.g. 60). Leave empty to use the multiplier.
Range 1 to 240
Qualitymci = best (moves detail along with motion, slowest). blend = fast cross-fade. dup = repeat frames.
Default: mci · Choices: dup, blend, mci
CrfCRF for libx264. Lower = higher quality/larger file. Default: 18
Default: 18 · Range 0 to 51

Tip: Prefer RIFE for general-purpose high-quality interpolation.

Time & Motion

Scrolling Pro

Apply video scrolling effect with definable directions.

Good for: VJ loops and projection mapping backgrounds · Infinite-scroll ambience for installations or screensavers · Graphic overlays behind text or UI elements

Settings

SettingWhat it does
Horizontal speedValue between -1.0 and 1.0 representing scroll speed left to right as a fraction of the frame size per frame. 0.01 scrolls left by 1 percent of the frame width per frame.
Default: 0.01 · Range -1 to 1 · Good start: -0.05–0.05
Vertical speedValue between -1.0 and 1.0 representing scroll speed up to down as a fraction of the frame size per frame. 0.005: Scrolls up by 0.5 percent of the frame height per frame.
Default: 0 · Range -1 to 1 · Good start: -0.05–0.05

Tip: Slight blur (e.g., via Blur Pix) before scrolling can remove hard seams in loop textures.

Time & Motion

Broken Scroll

Amplified frame differences plus a slow vertical scroll — a rolling, broken-tracking look.

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1
Time & Motion

Slit-Scan

Different parts of the frame show different moments in time — rows, rings, columns or waves of delay.

Settings

SettingWhat it does
PatternDefault: rows (time warp) · Choices: rows (time warp), radial, columns, wavy
BandsHow many strips/rings the frame is cut into.
Default: 48 · Range 8 to 200
Depth (frames)How far back in time the oldest band reaches.
Default: 48 · Range 4 to 120
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Time & Motion

Tunnel

The picture wrapped around the inside of a tunnel you fly down.

Settings

SettingWhat it does
SpeedNegative flies backwards.
Default: 90 · Range -400 to 400
Tiles aroundDefault: 3 · Range 1 to 8
Tiles deepDefault: 4 · Range 1 to 10
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Time & Motion

Little Planet

Polar-coordinate 'tiny planet' — the bottom of the picture becomes a small round world with sky all around.

Settings

SettingWhat it does
CurveLower = bigger planet.
Default: 0.8 · Range 0.3 to 2
Spin (°/sec)Default: 0 · Range -180 to 180
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Time & Motion

Strobe Hold

Stroboscope: hold each picture for N frames for a stuttering low-frame-rate look; optionally blink to a color for part of every hold, and vary the hold length randomly.

Settings

SettingWhat it does
Hold (frames)Each picture is held for this many frames.
Default: 4 · Range 1 to 60
Blink (%)How much of each hold shows the blink color instead of the picture (0 = never).
Default: 0 · Range 0 to 100
Blink colorBlink color. Default: #000000.
Default: #000000
Random holdHold lengths vary randomly from 1 up to Hold.
Time & Motion

Freeze Frame

Freeze frame: hold the picture still at a chosen moment. 'hold in place' replaces frames (length and audio unchanged); 'insert' adds the freeze, lengthening the clip, with silence under it.

Settings

SettingWhat it does
Freeze at (seconds)Time to freeze at, in seconds. Default: 2.
Default: 2 · Range 0 to 36000
Freeze for (seconds)How long to hold, in seconds. Default: 1.5.
Default: 1.5 · Range 0.04 to 600
Modehold in place = frames after the freeze point are replaced by the still. insert = the still is added, pushing the rest of the clip later (silence under it).
Default: hold in place (same length) · Choices: hold in place (same length), insert (clip gets longer)
Color & Look

Color Match

Color Match: move the clip's colour toward a reference clip or still (Reinhard LAB transfer). The colour is measured once over sampled frames, so the correction is the same across the whole clip.

Settings

SettingWhat it does
ReferenceConnect the clip or still whose colour you want to match. Connect a second video to its extra dot, or add one with + Input video.
Strength1 = the reference colour in full. Lower blends the match back toward the original.
Default: 1 · Range 0 to 1 · Good start: 0.3–1
Frames sampledFrames used to measure each clip's colour. More = steadier, slower to measure.
Default: 20 · Range 3 to 60 · Good start: 10–30
Color & Look

Local Contrast (CLAHE)

Local Contrast (CLAHE): equalise the contrast in small patches of the picture, so dark and bright areas both show detail. Only the lightness changes; the colours stay.

Settings

SettingWhat it does
Clip limitHow hard the contrast is pushed in each patch. Higher = more punch, and more grain.
Default: 2 · Range 0.5 to 8 · Good start: 1–4
Patch gridHow many patches the picture is split into across. More patches = more local.
Default: 8 · Range 2 to 32 · Good start: 4–16
Color & Look

Gamma Fix

Adjust gamma, brightness, contrast, and saturation.

Normalizes gamma, brightness, and exposure levels for broadcast-safe or web-safe consistency across diverse footage.

Good for: Normalizing multi-camera footage before editing · Preparing consistent brightness for LUT workflows or grading · Fixing underexposed or flat-looking archive material

Settings

SettingWhat it does
Target YavgTarget average luma (0..255). ~64 is a balanced web midpoint. Try 60–70 for darker footage, 70–90 for bright.
Default: 64 · Range 16 to 235
Min ContrastLower clamp for auto contrast mapping. Default 0.80.
Default: 0.8 · Range 0.5 to 1.5
Max ContrastUpper clamp for auto contrast mapping. Default 1.35.
Default: 1.35 · Range 1 to 3
GammaOptional gamma override (1.00 = neutral). Leave at 1.00 to rely on contrast/brightness mapping.
Default: 1 · Range 0.3 to 3
SatOptional saturation multiplier via hue filter (1.00 = unchanged). e.g., 1.10 = +10%% saturation.
Default: 1 · Range 0 to 3
LegalizeClamp output to broadcast-legal (TV) range using zscale and output yuv420p.
VcodecVideo codec for output. Default libx264.
Default: libx264
CrfCRF for output quality (x264/x265). Default 18.
Default: 18
PresetEncoder preset (x264/x265). Default medium.
Default: medium
AcodecAudio codec. Default aac.
Default: aac
AbAudio bitrate. Default 160k.
Default: 160k

Tip: Start with small gamma adjustments for natural tonal shifts.

Color & Look

LUT Apply

Apply a 3D LUT (.cube / .3dl) with optional colour adjustments.

Applies a 1D or 3D LUT to recolor a video, enabling film-style grading, color transforms, technical conversions, or creative look development.

Good for: Film emulation workflows and creative looks · Technical transforms (camera LOG → REC.709) · Experimenting with looks before importing into grading applications

Settings

SettingWhat it does
LUT File3D LUT file (.cube, .3dl)
InterpolationDefault: tetrahedral · Choices: tetrahedral, trilinear, nearest
LUT IntensityDefault: 1 · Range 0 to 1
BrightnessDefault: 0 · Range -1 to 1
ContrastDefault: 1 · Range 0 to 2
SaturationDefault: 1 · Range 0 to 3
GammaDefault: 1 · Range 0.1 to 10

Tip: Always preview LUTs with a variety of footage—some LUTs are extremely contrast-heavy.

Color & Look

Duotone

Maps luminance to a 2-color gradient (duotone) — a popular stylized-poster look. Shadows map to one color, highlights to another.

Settings

SettingWhat it does
Shadow ColorColor for the darkest areas. Default: #1a0038.
Default: #1a0038
Highlight ColorColor for the brightest areas. Default: #ffcc00.
Default: #ffcc00
Color & Look

Night Vision

Green-phosphor night-vision-goggle look — desaturated, tinted green, with sensor grain and lens vignette falloff.

Settings

SettingWhat it does
GrainGrain/noise amount. Default: 18.
Default: 18 · Range 0 to 60
BrightnessBrightness boost. Default: 0.05.
Default: 0.05 · Range -0.5 to 0.5
Color & Look

Old Film Damage

Vintage film-print damage — desaturated/faded vintage color grade, grain and dust, random brightness flicker, and a subtle gate-weave jitter.

Settings

SettingWhat it does
IntensityOverall strength multiplier. Default: 1.0.
Default: 1 · Range 0 to 1
Color & Look

Halftone

Newsprint-style halftone dot pattern — dot size in each cell is driven by the source's local brightness.

Settings

SettingWhat it does
Cell SizeSize in pixels of each dot cell. Smaller = finer detail, slower to render. Default: 8.
Default: 8 · Range 2 to 40
Color & Look

Steel Wash

Applies a cold blue steel filter to video.

Applies a cool steel-blue wash across the entire video image. The effect evokes the feel of industrial lighting, cold metal surfaces, or desaturated cinematic palettes often seen in sci-fi, dystopian, and neo-noir aesthetics.

Good for: Neo-noir or cyberpunk scenes · Cold, metallic grading for music videos · Industrial montage or mechanical close-ups

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Bad Contrast for harsher, high-impact steel tones.

Color & Look

Pickle Juice

As if the video was dipped in pickle juice.

” This look often includes green-tinted washes, acidic highlights, sour midtone shifts, and a brined, stained aesthetic reminiscent of aged film submerged in an odd chemical bath.

Good for: Stylized music videos seeking gritty, acidic toning · Collage-style edits where each clip has a unique chemical wash · VHS-simulation workflows combined with Bad Contrast or Nostalgic Stutter

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Pair with Gamma Fix before applying for more consistent pickle coloration across the image.

Color & Look

Septic

Makes the video look like a person in septic shock from drinking too much effluent.

Applies an intense, distressed visual filter inspired by the physiological collapse associated with septic shock.

Good for: Horror sequences and psychological thrillers · Music videos with themes of collapse, breakdown, or emotional overload · Glitch-art or experimental montage

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Bad Contrast for deeper shadows and harsher tonal decay.

Color & Look

WB Flare

Apply filter with a blown out white-balance flare.

Applies a blown-out white-balance flare that washes the frame in bright, overexposed warmth. The effect simulates a camera whose white balance has catastrophically failed, producing harsh brightness spikes, color melt, and an unstable exposure cast.

Good for: Music-video chorus drops or emotional surges · Stylized whiteouts in experimental film or collage · Dreamlike or transcendental sequences

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Pair with Overexposed Stutter for extreme blown-out instability.

Color & Look

WB Flare Pro

Apply filter with a blown out white-balance flare.

Settings

SettingWhat it does
Smoothing (sigma)Apply bilateral filter, spatial smoothing while preserving edgesSet sigma of gaussian function to calculate spatial weight. Allowed range is 0 to 512. Default is 0.1.
Default: 5 · Range 0.1 to 64 · Good start: 2–20
Color & Look

Xpiritualism

Multi-layer bloom with a pastel colour pass — soft, dreamy glow.

Settings

SettingWhat it does
StyleXpiritualism style preset: soft = gentle, pastel, minimal grain (default) deep = richer contrast, stronger vignette & bloom cosmic = more hue shift, grainier, dreamy
Default: soft · Choices: soft, deep, cosmic
Bloom RadiusBloom blur radius (luma_radius for boxblur). Higher = softer glow. Default: 8.0
Default: 8 · Range 0 to 40
Bloom StrengthBloom blend opacity (0.0–1.0). Default: 0.7
Default: 0.7 · Range 0 to 1
SaturationOverall saturation multiplier. Default: 1.15
Default: 1.15 · Range 0 to 3
Hue ShiftHue shift in DEGREES (used in ffmpeg hue filter). Default: 0.0
Default: 0 · Range -180 to 180
VignetteVignette blend strength (0.0–1.0-ish). Default: 0.35
Default: 0.35 · Range 0 to 1
GrainFilm grain strength (ffmpeg noise alls parameter). Default: 8.0
Default: 8 · Range 0 to 40
Color & Look

Zapruder

Simulates deteriorated or unstable historical footage.

Applies a stylized “Zapruder film” aesthetic inspired by the iconic 1960s 8mm footage. The effect simulates historical analog film qualities including jitter, grain, contrast instability, chromatic fading, and hand-held mechanical wobble.

Good for: Fake archival footage for narrative or documentary storytelling · Flashback sequences, memory montages, or historical reenactments · Vintage collage art and experimental cinema

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Add Blur Pix beforehand for softer, more analog edges.

Color & Look

Bad Predator

Apply bad Predator heat vision effect.

Applies a stylized “bad Predator heat vision” effect to the input video. This intentionally degraded thermal-vision look mimics the broken or low-budget interpretation of the thermal-view used in the Predator films — surreal color distortion, thermal-like banding, exaggerated edge detection, and unstable gradients.

Good for: Music videos, experimental film, and glitch-aesthetic sequences · Sci-fi projects referencing Predator-style heat vision · Stylized surveillance overlays or false-sensor aesthetics

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Apply before compression-heavy steps to preserve the color distortions.

Color & Look

Ball Point Pen

Frame-difference edges with deinterlace artifacts and colour balance — inked sketch look.

Applies a ball-point-pen illustration effect to the input video, simulating hand-drawn line textures, monochrome ink shading, and stylized pen-stroke contrast. The result evokes sketchbook drawings, technical notebook diagrams, or lo-fi comic-style renderings.

Good for: Rotoscope-style animation sequences · Comic-inspired music videos or narrative interludes · Experimental art filtering where realism is intentionally minimized

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Consider pairing with Convert Dims (square formats like 1080×1080) for Instagram-style sketch loops.

Color & Look

Light Snow

Motion-interpolated frames with chroma shift and debanding — a light static shimmer.

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1
Color & Look

RB Blur

Filter to reduce color banding in flat, gradient-heavy areas (e.g., skies, shadows).

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1
Color & Look

RB Blur Pro

Filter to reduce color banding in flat, gradient-heavy areas (e.g., skies, shadows).

Settings

SettingWhat it does
Debanding strengthDebanding intensity. Higher removes more banding but can soften fine detail. Try ~0.8–1.5; increase if banding persists.
Default: 1 · Range 0.51 to 4 · Good start: 0.8–1.5
Neighborhood (px)Debanding neighborhood size. Larger helps big smooth gradients (skies) but may smooth texture. Try 8–16.
Default: 12 · Range 4 to 32 · Good start: 8–16
Color & Look

Recalled Sensor

Causes a dramatic bloom or edge glow. Like overexposure/ video burn.

Simulates the look of a malfunctioning digital camera sensor that is defective, failing, or subject to an official recall. This effect introduces corrupted rows, unstable color channels, dead-pixel clusters, exposure tearing, and chaotic image instability — mimicking catastrophic sensor failure.

Good for: Horror, sci-fi, or techno-thriller sequences involving malfunctioning equipment · Glitch art and experimental cinema exploring digital decay · Surveillance or found-footage aesthetics that require broken-camera realism

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Pair with Bad Contrast for harsher tonal collapse.

Color & Look

Recalled Sensor Pro

Causes a dramatic bloom or edge glow. Like overexposure/ video burn.

Settings

SettingWhat it does
Glow radius (px)Glow neighborhood size for amplify. Small = tight edge glow; large = thicker bloom/halo. Try 2–8, push higher for heavy burn.
Default: 4 · Range 1 to 32 · Good start: 2–8
Glow intensityGlow intensity for amplify. Higher values increase bloom/burn and highlight clipping. Try 1.2–3.0; push higher for dramatic damage.
Default: 2 · Range 0.5 to 10 · Good start: 1.2–3

Presets: Soft glow Burn Meltdown

Color & Look

Soapblind

It makes the video look like you have soap in your eyes.

Applies a hazy, washed-over film that resembles the visual blur of “soap-blinded eyes” — a foggy, filmy distortion that softens detail and reduces clarity. The result is a dreamy, smeared, low-contrast haze that feels like staring through suds, steam, or a semi-opaque membrane.

Good for: Emotional, melancholic, or introspective scenes · Dream sequences or memory dissolves · Music videos needing soft glaze or atmospheric haze

Settings

SettingWhat it does
AmountHow much of the effect is mixed in. 1 = the full effect (default); 0 = the original picture.
Default: 1 · Range 0 to 1

Tip: Combine with Slight Smear for a double-soft aesthetic.

Color & Look

Dither

Adaptive-palette dithering: reduce the video to N colors picked from the footage and re-render it with a dithering algorithm. Use Pixel Size for chunky, retro dither.

Settings

SettingWhat it does
AlgorithmError-diffusion (floyd_steinberg, atkinson, sierra*, burkes, heckbert) or ordered (bayer).
Default: floyd_steinberg · Choices: floyd_steinberg, atkinson, sierra2, sierra2_4a, sierra3, burkes, heckbert, bayer
ColorsNumber of palette colors (2-256). Default: 16.
Default: 16 · Range 2 to 256
Bayer ScaleBayer pattern strength 0-5 (only used with the bayer algorithm). Default: 3.
Default: 3 · Range 0 to 5
Paletteglobal = one palette for the whole clip (stable). per_frame = a new palette every frame (shimmers).
Default: global · Choices: global, per_frame
Pixel SizeDither at 1/N resolution and scale up for chunky pixels (1 = full resolution). Default: 2.
Default: 2 · Range 1 to 16
Crfx264 quality (lower = better). Dither patterns need a low CRF to survive compression. Default: 14.
Default: 14 · Range 0 to 51
Color & Look

Retro Dither

Dither onto a fixed palette: 1-bit B&W, Game Boy, CGA, EGA, C64, PICO-8, phosphor monitors, or your own colors. Tweak Contrast/Brightness first for best 1-bit results.

Settings

SettingWhat it does
PalettePick a classic palette, or 'custom' to use the hex list below.
Default: bw · Choices: bw, gameboy, cga_pink, cga_green, ega16, c64, pico8, amber, green_phosphor, sepia, virtualboy, custom
Custom colorsComma-separated hex colors for the 'custom' palette, e.g. #000000, #ff6847, #ffffff.
AlgorithmDithering algorithm. Default: floyd_steinberg.
Default: floyd_steinberg · Choices: floyd_steinberg, atkinson, sierra2, sierra2_4a, sierra3, burkes, heckbert, bayer
Bayer ScaleBayer pattern strength 0-5 (only with the bayer algorithm). Default: 3.
Default: 3 · Range 0 to 5
Pixel SizeDither at 1/N resolution and scale up for chunky pixels (1 = full resolution). Default: 2.
Default: 2 · Range 1 to 16
ContrastContrast before dithering (1.0 = unchanged). Default: 1.2.
Default: 1.2 · Range 0.5 to 2
BrightnessBrightness before dithering, -1 to 1. Default: 0.
Default: 0 · Range -0.5 to 0.5
Crfx264 quality (lower = better). Default: 14.
Default: 14 · Range 0 to 51
Color & Look

Ordered Dither

Bayer, clustered-dot, blue-noise and static dither patterns onto a palette or posterized colors, optionally animated.

Settings

SettingWhat it does
PatternbayerN = classic ordered grid, clustered = halftone-like dots, blue_noise / ign = smooth film-like noise, white_noise = harsh static.
Default: bayer8 · Choices: bayer2, bayer4, bayer8, bayer16, clustered, blue_noise, ign, white_noise
Palette'levels' posterizes each color channel to N steps; others snap to a named palette.
Default: levels · Choices: levels, bw, gameboy, cga_pink, cga_green, ega16, c64, pico8, amber, green_phosphor, sepia, virtualboy, custom
Custom colorsComma-separated hex colors for the 'custom' palette.
LevelsSteps per color channel when --palette levels (2 = 8 colors). Default: 2.
Default: 2 · Range 2 to 16
StrengthDither amount: 0 = hard quantize, 1 = full, >1 = noisier. Default: 1.
Default: 1 · Range 0 to 1
Pixel SizeDither at 1/N resolution and scale up for chunky pixels. Default: 2.
Default: 2 · Range 1 to 16
AnimateChange the pattern every frame (boiling/flicker look).
Crfx264 quality (lower = better). Default: 14.
Default: 14 · Range 0 to 51
Color & Look

Neon Edges

Neon edge outlines: detect edges (Canny, Laplacian or Sobel) and draw them as glowing colored lines over a dimmed, original or black background.

Settings

SettingWhat it does
Edge methodEdge detector. Default: canny.
Default: canny · Choices: canny, laplacian, sobel
Line colorLine color. Default: #00FFFF.
Default: #00FFFF
RainbowCycle the line color through the rainbow over time.
SensitivityHow many edges are picked up, 0 (few) to 1 (many). Default: 0.7.
Default: 0.7 · Range 0 to 1
ThicknessLine thickness in px. Default: 2.
Default: 2 · Range 1 to 10
GlowGlow radius in px (0 = no glow). Default: 12.
Default: 12 · Range 0 to 50
BackgroundWhat shows behind the neon lines.
Default: dimmed · Choices: dimmed, original, black
DimBrightness of the dimmed background, 0-1. Default: 0.25.
Default: 0.25 · Range 0 to 1
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Color & Look

Cartoon

Cartoon look: edge-preserving smoothing, posterized flat colors and black ink outlines from adaptive thresholding.

Settings

SettingWhat it does
LevelsColor steps per channel (2-16). Default: 6.
Default: 6 · Range 2 to 16
SmoothSmoothing passes — more = flatter, more painted. Default: 3.
Default: 3 · Range 1 to 15
Edge SizeOutline detail window, odd number: small = fine lines, large = bold. Default: 9.
Default: 9 · Range 3 to 21
Edge StrengthOutline threshold offset — higher = fewer lines. Default: 4.
Default: 4 · Range 1 to 15
SaturationColor boost. Default: 1.25.
Default: 1.25 · Range 0 to 3
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Color & Look

Sketch

Sketch and painterly looks (OpenCV photo filters): pencil drawing, colored pencil, stylization, edge-preserving paint smoothing, detail enhancement. Processes at reduced size for speed.

Settings

SettingWhat it does
Modepencil = grey pencil drawing, color_pencil = tinted pencil, stylize = watercolor-ish, smooth_paint = edge-preserving flatten, detail = crisp detail boost.
Default: pencil · Choices: pencil, color_pencil, stylize, smooth_paint, detail
SpatialSmoothing size (sigma_s, 0-200). Default: 60.
Default: 60 · Range 10 to 200
RangeEdge sensitivity (sigma_r, 0-1): low keeps more detail. Default: 0.3.
Default: 0.3 · Range 0 to 1
ShadePencil shading darkness (0-0.1). Default: 0.05.
Default: 0.05 · Range 0 to 0.2
Work ScaleProcess at this fraction of the frame size, then scale back up (speed). Default: 0.5.
Default: 0.5 · Range 0.1 to 1
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Color & Look

K-Means Palette

Snap the video to its N dominant colors, optionally dithered.

Settings

SettingWhat it does
ColorsNumber of palette colors (2-32). Default: 6.
Default: 6 · Range 2 to 32
RefreshRecompute the palette every N frames. Default: 24.
Default: 24 · Range 1 to 120
DitherOptionally dither between the palette colors instead of hard banding.
Default: none · Choices: none, bayer4, bayer8, blue_noise, ign
Pixel SizeChunky pixels: process at 1/N size. Default: 1.
Default: 1 · Range 1 to 16
Crfx264 quality. Default: 16.
Default: 16 · Range 0 to 51
Color & Look

Filter Library

Every photo-booth filter in one list — plus any you drop into ~/.videobeaux/filters/ (Negative, Game Boy, VHS, CRT, halftone, comic, fisheye…).

Settings

SettingWhat it does
FilterGrouped by pack (Classic, Retro TV, Print & pixel, Art, Colour, Warp, Face, Time…).
Default: Classic · Negative · Choices: Classic · Negative, Classic · Black & white, Classic · Thermal, Classic · Pop art, Classic · Sketch, Classic · Pixelate, Classic · Glitch, Classic · Kaleidoscope, Classic · Sepia, Classic · Night vision, Slit-scan · Time warp, Slit-scan · Radial, Slit-scan · Vertical, Slit-scan · Wavy, Time · Ghost trails, Time · RGB time-split, Time · Motion only, Retro TV · VHS, Retro TV · CRT, Retro TV · Weak signal, Print & pixel · Halftone, Print & pixel · Game Boy, Print & pixel · 1-bit Mac, Print & pixel · ASCII, Print & pixel · LED wall, Print & pixel · CGA, Print & pixel · Commodore 64, Print & pixel · PICO-8, Art · Comic, Art · Oil paint, Art · Watercolour, Art · Neon edges, Art · Emboss, Art · Blueprint, Colour · Duotone cyan/magenta, Colour · Duotone orange/teal, Colour · Duotone purple/yellow, Colour · Infrared, Colour · Solarize, Colour · Posterize, Colour · Chromatic aberration, Colour · Lomo, Colour · Hue cycle, Warp · Fisheye, Warp · Pinch, Warp · Swirl, Warp · Wavy mirror, Warp · Tunnel, Warp · Little planet, Face · Big head, Face · Tiny head, Face · Big eyes, Face · Face swap
Amount1 = full filter, lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The filter is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

ASCII Art

ASCII art video: replaces each block of the picture with a text character matching its brightness. Pick a character set, colors (source / mono / matrix green / amber), size and an edge boost.

Settings

SettingWhat it does
CharactersWhich characters build the picture. "blocks" uses shaded squares; "custom" uses the text below.
Default: classic · Choices: classic, detailed, simple, dots, binary, blocks, custom
Custom charactersUsed when Characters = custom. Order doesn't matter — they're sorted by how much ink each uses.
ColorColor style.
Default: source colors · Choices: source colors, mono, matrix green, amber terminal
Text color (mono)Text color for mono. Default: #FFFFFF.
Default: #FFFFFF
Background (mono)Background color for mono. Default: #000000.
Default: #000000
Characters acrossMore = finer detail, smaller characters.
Default: 110 · Range 20 to 400
Edge boostMakes outlines pop (0 = off).
Default: 0 · Range 0 to 1
InvertFlip the brightness mapping: dark areas get the dense characters (pair with a light background color in mono mode).
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Color & Look

Negative

Invert the picture like a photo negative — or flip only the brightness and keep the colors.

Settings

SettingWhat it does
Invertfull = every color inverted. brightness only = light/dark flipped, hues kept.
Default: full · Choices: full, brightness only
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Black & White

High-contrast black and white with local contrast boost (faces and texture pop), optional film grain.

Settings

SettingWhat it does
ContrastLocal contrast boost.
Default: 2.5 · Range 0.5 to 8
GrainFilm grain amount.
Default: 0 · Range 0 to 1
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Sepia & Tones

Antique single-tone looks: sepia, cyanotype blue, rose, forest or gold.

Settings

SettingWhat it does
ToneDefault: sepia · Choices: sepia, cyanotype, rose, forest, gold
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Thermal

Thermal-camera false color — pick the heat palette (inferno, jet, turbo, hot, plasma…).

Settings

SettingWhat it does
PaletteDefault: inferno · Choices: inferno, jet, turbo, hot, magma, plasma, viridis, rainbow, ocean, bone
Skip auto-contrastDon't stretch the brightness range first.
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Infrared Film

False-color infrared film look — foliage turns pink and red, skies go dark.

Settings

SettingWhat it does
SaturationDefault: 1.5 · Range 0 to 3
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Solarize

Darkroom solarization — tones above the threshold flip, giving glowing metallic edges.

Settings

SettingWhat it does
ThresholdBrightness where the tones fold over.
Default: 128 · Range 1 to 254
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Posterize

Reduce each color channel to a few flat levels.

Settings

SettingWhat it does
LevelsShades per channel.
Default: 4 · Range 2 to 16
Keep edges sharpSkip the slight pre-blur.
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Lomo

Cross-processed toy-camera look: punchy curves, color cast and dark vignette corners.

Settings

SettingWhat it does
VignetteCorner darkening.
Default: 0.65 · Range 0 to 1.2
SaturationDefault: 1 · Range 0 to 2.5
Curve strengthDefault: 1 · Range 0.4 to 2
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Hue Cycle

Rotate every color around the color wheel continuously.

Settings

SettingWhat it does
Speed (°/sec)Negative runs the wheel backwards.
Default: 60 · Range -360 to 360
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Pop Art

Warhol-style flat-color panels — four colorways in a 2×2 grid, or one palette over the whole frame.

Settings

SettingWhat it does
LayoutDefault: 2x2 panels · Choices: 2x2 panels, single
Palette (single)Which colorway when Layout = single (0-3).
Default: 0 · Range 0 to 3
Shadow cutBrightness where dark turns to mid tone.
Default: 90 · Range 0 to 254
Highlight cutBrightness where mid turns to light.
Default: 170 · Range 1 to 255
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Blueprint

Technical-drawing look — white edge lines on blueprint blue, with an optional grid.

Settings

SettingWhat it does
Line detailMore lines at higher values.
Default: 0.5 · Range 0 to 1
Line colorDefault: #EBF5FF
Paper colorDefault: #0A4696
Draw grid
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Emboss

Raised-relief emboss lit from any angle, in gray or keeping the colors.

Settings

SettingWhat it does
StrengthDefault: 1 · Range 0 to 4
Light angle (°)Default: 135 · Range 0 to 360
Keep color
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Oil Paint

Painterly oil-paint look: smoothed brush regions, posterized tones and a little canvas relief.

Settings

SettingWhat it does
Brush sizeLarger = broader strokes.
Default: 40 · Range 5 to 100
Tone stepLarger = flatter color regions.
Default: 24 · Range 2 to 64
ReliefDefault: 0.35 · Range 0 to 1
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

LED Wall

The picture rebuilt from a grid of round glowing LEDs, like a stadium video wall.

Settings

SettingWhat it does
LED size (px)Default: 10 · Range 3 to 60
BrightnessDefault: 1.15 · Range 0.5 to 2
Dot sizeDefault: 1 · Range 0.3 to 1.3
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Pixelate

Chunky square pixels — a mosaic of any block size.

Settings

SettingWhat it does
Block size (px)Default: 16 · Range 2 to 200
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Color & Look

Color Pass

Color pass: keep one color range and turn everything else gray (or remove just that color). Pick the color, then widen or narrow the range and softness.

Settings

SettingWhat it does
ColorPick the color to keep (or remove); only its hue matters.
Default: #FF0000
ModeKeep or remove the color.
Default: keep this color · Choices: keep this color, remove this color
Range (°)How far around the color wheel still counts as the color.
Default: 30 · Range 1 to 180
Softness (°)Fade at the edge of the range.
Default: 20 · Range 0 to 90
Ignore grays belowPixels less colorful than this are treated as gray, never as the picked color.
Default: 0.15 · Range 0 to 1
Drain the rest1 = fully gray, lower keeps some color.
Default: 1 · Range 0 to 1
Boost the colorExtra saturation on the kept color.
Default: 1 · Range 0 to 3
Color & Look

Proc Amp

Proc amp: brightness, contrast, saturation, hue rotation, gamma, black/white levels and color temperature, with an optional broadcast-safe clamp. Audio is copied untouched.

Settings

SettingWhat it does
Brightness0 = unchanged.
Default: 0 · Range -1 to 1
Contrast1 = unchanged.
Default: 1 · Range 0 to 3
Saturation1 = unchanged, 0 = black and white.
Default: 1 · Range 0 to 3
Hue rotate (°)Hue rotation in degrees. Default: 0.
Default: 0 · Range -180 to 180
Gamma1 = unchanged; above 1 lifts the mid-tones.
Default: 1 · Range 0.1 to 3
Black levelEverything at or below this becomes black.
Default: 0 · Range 0 to 0.5
White levelEverything at or above this becomes white.
Default: 1 · Range 0.5 to 1
Color temperature (K)6500 = unchanged. Lower is warmer (orange), higher is cooler (blue).
Default: 6500 · Range 2000 to 12000
Broadcast-safeClamp to legal video levels (16-235 luma).
Vision & Tracking

Motion Mask

Shows what moves: a black-and-white mask, the picture with the moving parts tinted, or only the moving parts (background subtraction).

Settings

SettingWhat it does
Lookmask = white where things move. highlight = the picture with the moving parts tinted red. isolate = only the moving parts, on black.
Default: mask · Choices: mask, highlight, isolate
Memory (frames)How many frames the background is learned from. Longer = slower to forget something that stopped moving.
Default: 200 · Range 10 to 1000 · Good start: 60–400
SensitivityHow different a pixel must be from the background to count as moving. Lower = more sensitive.
Default: 16 · Range 4 to 100 · Good start: 10–40
Invert (show what stands still)Swap the moving and still parts.
Vision & Tracking

Face Track

Detect and track faces and draw tracking graphics over them: boxes, corner brackets, circles, crosshairs, IDs, motion trails, a spotlight, or an image pasted on each face.

Settings

SettingWhat it does
StyleTracking graphic drawn on each face. 'none' draws nothing (use with Spotlight or Image).
Default: corners · Choices: corners, box, circle, crosshair, none
ColorGraphic color. Default: #00FF88.
Default: #00FF88
ThicknessLine thickness in px. Default: 3.
Default: 3 · Range 1 to 12
PaddingGrow the face box by this fraction before drawing. Default: 0.15.
Default: 0.15 · Range 0 to 1
Show IdLabel each face with its track ID.
TrailDraw a motion trail behind each face.
SpotlightDarken everything outside the faces by this amount, 0-1. Default: 0 (off).
Default: 0 · Range 0 to 1
Overlay image (optional)PNG/JPG pasted onto each face, scaled to fit. PNGs with transparency work best.
Detect EveryRun face detection every N frames (tracking fills the gaps). Default: 3.
Default: 3 · Range 1 to 30
Min ConfidenceDetection confidence threshold, 0-1. Default: 0.6.
Default: 0.6 · Range 0 to 1
SmoothingTrack smoothing, 0 = snappy, 0.9 = very smooth. Default: 0.5.
Default: 0.5 · Range 0 to 1
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Vision & Tracking

Face Redact

Obscure tracked faces with blur, pixelation, a solid fill or 1-bit dither. Use Invert to hide everything except the faces.

Settings

SettingWhat it does
EffectObscuring effect. Default: blur.
Default: blur · Choices: blur, pixelate, solid, dither
ShapeRegion shape. Default: ellipse.
Default: ellipse · Choices: ellipse, box
StrengthBlur radius / pixel block size, relative to face size (1-100). Default: 30.
Default: 30 · Range 1 to 100
Fill color (solid)Fill color for the solid effect.
Default: #000000
PaddingGrow the face region by this fraction (covers hair/ears and tracking lag). Default: 0.25.
Default: 0.25 · Range 0 to 1
InvertObscure everything except the faces.
Detect EveryRun face detection every N frames. Lower = safer for fast motion. Default: 2.
Default: 2 · Range 1 to 30
Min ConfidenceDetection confidence threshold, 0-1. Lower catches more faces. Default: 0.5.
Default: 0.5 · Range 0 to 1
SmoothingTrack smoothing, 0 = snappy. Keep low for redaction. Default: 0.3.
Default: 0.3 · Range 0 to 1
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Vision & Tracking

Face Follow

Smart reframe: a smoothed virtual camera that pans and zooms to keep a face in frame. Output keeps the input's size.

Settings

SettingWhat it does
FollowWhich face to follow when several are visible.
Default: largest · Choices: largest, first, center
ZoomFixed zoom factor (>=1). Default: 1.6.
Default: 1.6 · Range 1 to 4
Auto ZoomPick the zoom so the face fills --face_fraction of the frame width (ignores --zoom).
Face FractionTarget face width as a fraction of frame width with --auto_zoom. Default: 0.3.
Default: 0.3 · Range 0.05 to 0.9
Face YWhere the face sits vertically in the frame, 0 = top, 1 = bottom. Default: 0.42.
Default: 0.42 · Range 0 to 1
SmoothingCamera smoothness, 0 = locked on, 0.99 = lazy drift. Default: 0.9.
Default: 0.9 · Range 0 to 1
When no facehold = stay where the camera was; zoom_out = ease back to the full frame.
Default: hold · Choices: hold, zoom_out
Detect EveryRun detection every N frames. Default: 3.
Default: 3 · Range 1 to 30
Min ConfidenceDetection confidence threshold. Default: 0.6.
Default: 0.6 · Range 0 to 1
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Vision & Tracking

Motion Ghost

Motion isolation with background subtraction (MOG2): tint what moves, show only the moving parts, or leave glowing ghost trails behind them. Works best with a still camera.

Settings

SettingWhat it does
Modehighlight = tint movers over a dimmed scene, silhouette = only the moving areas, trails = glowing ghost trails of motion.
Default: highlight · Choices: highlight, silhouette, trails
Tint colorTint/trail color. Default: #FF2D95.
Default: #FF2D95
Sensitivity0 = only big movement, 1 = pick up subtle movement. Default: 0.5.
Default: 0.5 · Range 0 to 1
DecayTrails: how slowly trails fade, 0-0.99. Default: 0.92.
Default: 0.92 · Range 0 to 1
DimBrightness of the non-moving scene, 0-1. Default: 0.35.
Default: 0.35 · Range 0 to 1
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Vision & Tracking

Feature Trails

Tracking-HUD look: finds corner features, follows them with optical flow and draws points, motion trails and connecting lines between nearby points.

Settings

SettingWhat it does
PointsMaximum tracked points. Default: 140.
Default: 140 · Range 10 to 500
Trail LengthFrames of trail behind each point. Default: 10.
Default: 10 · Range 2 to 60
Link DistanceConnect points closer than this fraction of the frame width (0 = no lines). Default: 0.12.
Default: 0.12 · Range 0 to 0.5
ColorGraphic color. Default: #7CFFCB.
Default: #7CFFCB
BackgroundBackground. Default: dimmed.
Default: dimmed · Choices: dimmed, original, black
DimBrightness of the dimmed background, 0-1. Default: 0.45.
Default: 0.45 · Range 0 to 1
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Vision & Tracking

Face Warp

Big head, tiny head or big eyes — warps tracked faces (works on several faces at once).

Settings

SettingWhat it does
EffectDefault: big head · Choices: big head, tiny head, big eyes
StrengthDefault: 1 · Range 0 to 2
SizeArea of the warp.
Default: 1 · Range 0.5 to 2
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Vision & Tracking

Face Swap

Swap two faces in the shot, copy the biggest face onto the others, or put a photo's face on everyone — aligned on the eyes and mouth and colour-matched to the skin.

Settings

SettingWhat it does
ModeSwap two faces needs two people in the same frame · Biggest face onto the others copies the largest face to every other face · Photo onto every face works with just one person (choose a Face photo).
Default: Swap two faces · Choices: Swap two faces, Biggest face onto the others, Photo onto every face
Face photoA picture of a face (JPEG/PNG). Used by "Photo onto every face".
Edge softnessHow gently the pasted face blends into its surroundings.
Default: 0.06 · Range 0 to 0.3
Skin-tone match0 = keep the pasted face's own colours · 1 = fully recoloured to the skin it lands on.
Default: 0.85 · Range 0 to 1
Amount1 = full effect; lower values blend the original back in.
Default: 1 · Range 0 to 1
Detail (px)The effect is applied at this size on the short side, then scaled back up — keeps the look consistent on any resolution and is faster. 0 = full resolution.
Default: 720 · Range 0 to 2160
Layout & Overlay

Clean Plate

Removes people and cars that move through a locked-off shot: the empty scene is built from the median of sampled frames, then the moving parts are replaced with it.

Settings

SettingWhat it does
Moderemove_movers = take the empty scene wherever something moves through the shot. plate_only = the empty scene itself, for the whole clip.
Default: remove_movers · Choices: remove_movers, plate_only
SensitivityHow different a pixel must be from the empty scene to count as something moving. Lower = more sensitive.
Default: 28 · Range 4 to 120 · Good start: 15–50
Frames sampled for the plateMore frames = a cleaner plate, but a slower first pass.
Default: 25 · Range 5 to 80 · Good start: 15–40
Processing width (px)The plate and the moving-part mask are worked out at this width, then scaled up.
Default: 960 · Range 160 to 1920 · Good start: 480–960
Layout & Overlay

Key Composite

Key Composite: put the subject over a new background, using a matte video (white = keep, black = remove). Connect a Remove Background node set to View = matte to its Matte input.

Settings

SettingWhat it does
Matte videoConnect a Remove Background node set to View = matte (or any grayscale video): white = keep, black = remove. Connect a second video to its extra dot, or add one with + Input video.
BackgroundWhat shows behind the subject. auto = the connected/picked Background video, else the Background image, else the Background color. transparent = no background at all: save as .webm or .mov (Output node format).
Default: auto · Choices: auto, color, image, video, transparent
Background colorBackground color. Default: #000000.
Default: #000000
Background imageBackground image.
Background videoConnect a video here (or browse) to put the subject on top of it. Loops if shorter. Connect a second video to its extra dot, or add one with + Input video.
ViewUse matte / checkerboard while tuning: white = kept, black = removed. matte gives a grayscale video you can connect to Key Composite.
Default: composite · Choices: composite, matte, checkerboard
AudioWhich sound the result keeps. Background video / Mix need a Background video connected.
Default: Foreground video · Choices: Foreground video, Background video, Mix both, Silent
Edge shrinkEat into the edge of the subject to lose fringing (pixels).
Default: 1 · Range 0 to 10
Edge featherSoften the edge of the matte (pixels).
Default: 1 · Range 0 to 10
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Layout & Overlay

Audio Visualizer

Draw the sound over the picture: a waveform line (wave) or a spectrum of bars (bars), in a strip at the bottom or top of the frame, or across the whole picture. The sound itself is kept.

Settings

SettingWhat it does
Stylewave = a line that follows the sound. bars = a spectrum, one bar per band of pitches.
Default: wave · Choices: wave, bars
Positionbottom or top = a strip of the chosen height. full = drawn over the whole picture.
Default: bottom · Choices: bottom, top, full
Strip height (% of frame)How tall the strip is, as a share of the picture. Ignored for full.
Default: 20 · Range 5 to 100 · Good start: 10–40
ColorColour of the line or bars.
Default: #FFFFFF
Opacity1 = solid. Lower lets the picture show through the drawing.
Default: 0.9 · Range 0 to 1 · Good start: 0.3–1
Layout & Overlay

Ken Burns

A slow push-in or pull-out across the whole clip, closing in on a focus point you choose. Works on stills and footage.

Settings

SettingWhat it does
Start zoomZoom at the first frame. 1 = the whole picture.
Default: 1 · Range 1 to 4 · Good start: 1–2
End zoomZoom at the last frame. Bigger than the start zooms in; smaller pulls out.
Default: 1.3 · Range 1 to 4 · Good start: 1–2
Focus across (0 left, 1 right)The point the zoom closes in on, horizontally. 0.5 = the middle.
Default: 0.5 · Range 0 to 1 · Good start: 0–1
Focus down (0 top, 1 bottom)The point the zoom closes in on, vertically. 0.5 = the middle.
Default: 0.5 · Range 0 to 1 · Good start: 0–1
Layout & Overlay

Watermark / Image Overlay

Overlay a watermark or image onto the video — 9-point placement or custom X/Y, scale or exact pixel sizing, opacity, spin, and a timed enable window.

Applies image or text watermarks onto video with configurable positioning, scaling, opacity, and blend style.

Good for: Artist signatures, branding marks, portfolio reels · Subtle watermarks for social media videos · Bold center overlays for drafts, screeners, and pre-release content

Settings

SettingWhat it does
WatermarkPath to watermark/overlay image (PNG/JPG/GIF).
PlacementWhere to position the overlay. Choose 'custom' to use X/Y Position below for exact pixel/expression control.
Default: bottom-right · Choices: top-left, top-center, top-right, center-left, center, center-right, bottom-left, bottom-center, bottom-right, custom
MarginMargin (px) from edges for placement.
Default: 24 · Range 0 to 500
X PosHorizontal position (pixels or an ffmpeg expression). Only used when Placement is 'custom'.
Y PosVertical position (pixels or an ffmpeg expression). Only used when Placement is 'custom'.
ScaleScale factor relative to watermark intrinsic width (iw*scale). Ignored if --width/--height is set.
Default: 0.25 · Range 0.02 to 1
WidthExact overlay width in pixels. Overrides --scale when set.
HeightExact overlay height in pixels (or -1 to preserve aspect ratio). Overrides --scale when set.
OpacityWatermark opacity (0.0–1.0).
Default: 0.8 · Range 0 to 1
SpinWatermark spin in degrees per second (0 = no rotation).
Default: 0 · Range -360 to 360
Wobble (°)Swings the image this many degrees one way, back through upright, then the same amount the other way. 0 = none.
Default: 0 · Range 0 to 180
Wobble Period (s)Seconds for one full wobble: out, back through upright, the other way, and back. Only used with Wobble.
Default: 2 · Range 0.1 to 60
Appear atWhen the image appears: seconds (83.5) or a timestamp (1:23, 0:01:23.5). 0 = from the start.
Default: 0
Disappear atWhen the image goes away: seconds or a timestamp. 0 = it stays to the end.
Default: 0
Wm LoopAdditional loops for watermark input (-1=infinite, 0=none, N>0 times).
Default: 0 · Range 0 to 50
Ignore LoopFor GIF watermark: ignore intrinsic loop (use frames once).
Video CrfCRF for libx264.
Default: 18 · Range 0 to 51
Video Presetx264 preset.
Default: fast

Tip: Keep watermark assets at 2× resolution for sharp results after scaling.

Layout & Overlay

Text Overlay

Overlay a block of text (a title, a lower third, a caption card) on the video. Same font and colour options as Auto Narrate; the same placement grid and Appear/Disappear timing as Watermark.

Settings

SettingWhat it does
TextThe words to show. Press Enter for a second line.
FontFont name (or a path to a .ttf/.otf file)
Default: Arial
Font SizeFont size in px. Every title uses this size; long text wraps onto more lines instead of shrinking
Default: 76
Text ColorColour of the text
Default: #FFFFFF
Outline ColorText outline colour
Default: #000000
PlacementWhere the text sits. Choose 'custom' to use X/Y Position below
Default: bottom-center · Choices: top-left, top-center, top-right, center-left, center, center-right, bottom-left, bottom-center, bottom-right, custom
Margin (px)Distance from the edges for the placement grid
Default: 24 · Range 0 to 500
X PositionHorizontal position in pixels or an ffmpeg expression. Only used when Placement is 'custom'
Y PositionVertical position in pixels or an ffmpeg expression. Only used when Placement is 'custom'
OpacityText opacity, 0 to 1
Default: 1 · Range 0 to 1
Spin (°/s)Continuous rotation in degrees per second, like Watermark. 0 = none
Default: 0 · Range -360 to 360
Wobble (°)Swings the text this many degrees one way, back through upright, then the same amount the other way. 0 = none
Default: 0 · Range 0 to 180
Wobble Period (s)Seconds for one full wobble: out, back through upright, the other way, and back. Only used with Wobble
Default: 2 · Range 0.1 to 60
Appear atWhen the text appears: seconds (83.5) or a timestamp (1:23). 0 = from the start
Default: 0
Disappear atWhen the text goes away: seconds or a timestamp. 0 = it stays to the end
Default: 0
Layout & Overlay

Remove Background

Cut the subject out and put anything behind it — or export transparent WebM/MOV. Static-camera mode needs no download; optional local AI models for harder shots.

Settings

SettingWhat it does
Enginestatic camera works instantly if the camera is locked off. The AI engines need a one-time model download from Setup → Local AI features.
Default: static camera (no download) · Choices: static camera (no download), AI · fast (u2netp), AI · people (u2net human)
Sensitivity (static)Static engine: how different from the empty scene a pixel must be to count as subject. Lower = more.
Default: 28 · Range 2 to 120
Threshold (AI)AI engines: matte level that counts as subject.
Default: 0.5 · Range 0 to 1
Edge softnessEdge softness of the matte, 0 to 1. Default: 0.25.
Default: 0.25 · Range 0.02 to 1
Temporal smoothingBlends each frame's matte with the previous one to stop flicker.
Default: 0.5 · Range 0 to 0.95
Fill holesFill gaps inside the subject.
Keep largest subject onlyDrop small stray blobs; keep the biggest connected shape.
InvertKeep the background instead of the subject.
Processing width (px)The matte is computed at this width, then scaled up.
Default: 960 · Range 240 to 3840
BackgroundWhat shows behind the subject. auto = the connected/picked Background video, else the Background image, else the Background color. transparent = no background at all: save as .webm or .mov (Output node format).
Default: auto · Choices: auto, color, image, video, transparent
Background colorBackground color. Default: #000000.
Default: #000000
Background imageBackground image.
Background videoConnect a video here (or browse) to put the subject on top of it. Loops if shorter. Connect a second video to its extra dot, or add one with + Input video.
ViewUse matte / checkerboard while tuning: white = kept, black = removed. matte gives a grayscale video you can connect to Key Composite.
Default: composite · Choices: composite, matte, checkerboard
AudioWhich sound the result keeps. Background video / Mix need a Background video connected.
Default: Foreground video · Choices: Foreground video, Background video, Mix both, Silent
Edge shrinkEdge shrink in pixels (0-10). Default: 1.
Default: 1 · Range 0 to 10
Edge featherEdge feather in pixels (0-10). Default: 1.
Default: 1 · Range 0 to 10
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Layout & Overlay

Chroma Key

Remove a green/blue screen (or any solid color) and put a color, image or another video behind — or export transparent WebM/MOV. Auto-detects the screen color.

Settings

SettingWhat it does
Key colorThe screen color to remove. Typical green screen ≈ #00B140, blue ≈ #0047BB.
Default: #00B140
Auto-detect key colorSample the edges of the first frame to find the screen color (ignores the color picker).
MethodKeying method.
Default: chromakey (YUV, best for most) · Choices: chromakey (YUV, best for most), colorkey (RGB)
SimilarityHow far from the key color still counts as screen. Raise until the background is gone.
Default: 0.15 · Range 0.01 to 1
Edge softnessGradual transparency at the edge of the key.
Default: 0.05 · Range 0 to 1
Spill removalRemoves the green/blue tint reflected onto the subject. 0 = off.
Default: 0.3 · Range 0 to 1
BackgroundWhat shows behind the subject. auto = the connected/picked Background video, else the Background image, else the Background color. transparent = no background at all: save as .webm or .mov (Output node format).
Default: auto · Choices: auto, color, image, video, transparent
Background colorBackground color. Default: #000000.
Default: #000000
Background imageBackground image.
Background videoConnect a video here (or browse) to put the subject on top of it. Loops if shorter. Connect a second video to its extra dot, or add one with + Input video.
Viewcheckerboard = see what was removed (start here) · matte = white kept / black removed · composite = the final result over your Background.
Default: checkerboard · Choices: composite, matte, checkerboard
AudioWhich sound the result keeps. Background video / Mix need a Background video connected.
Default: Foreground video · Choices: Foreground video, Background video, Mix both, Silent
Edge shrinkEat into the edge of the subject to lose fringing (pixels).
Default: 1 · Range 0 to 10
Edge featherSoften the edge of the matte (pixels).
Default: 1 · Range 0 to 10
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51

Presets: Green screen Blue screen Find the screen for me Tight edges (cleaner) Final: over background Export with transparency

Layout & Overlay

Luma Key

Luma key: make the darkest (or brightest) parts transparent and put something behind the rest. Choose Composite = screen / add for fire, smoke and light-leak overlays on black.

Settings

SettingWhat it does
Removedark = knock out blacks · bright = knock out whites/sky · custom = knock out around Threshold.
Default: dark · Choices: dark, bright, custom
Threshold (custom)Brightness to remove for Remove = custom (0 black … 1 white).
Default: 0.5 · Range 0 to 1
ToleranceHow wide a brightness range gets removed.
Default: 0.12 · Range 0 to 1
Edge softnessGradual fade at the edge of the range.
Default: 0.08 · Range 0 to 1
Invert matteKeep what would be removed, remove what would be kept.
Compositekey = cut out with transparency. screen / add / lighten = blend onto the background (best for fire, smoke, light on black).
Default: key (cut out) · Choices: key (cut out), screen, add, lighten
BackgroundWhat shows behind the subject. auto = the connected/picked Background video, else the Background image, else the Background color. transparent = no background at all: save as .webm or .mov (Output node format).
Default: auto · Choices: auto, color, image, video, transparent
Background colorBackground color. Default: #000000.
Default: #000000
Background imageBackground image.
Background videoConnect a video here (or browse) to put the subject on top of it. Loops if shorter. Connect a second video to its extra dot, or add one with + Input video.
Viewcheckerboard = see what was removed (start here) · matte = white kept / black removed · composite = the final result over your Background.
Default: checkerboard · Choices: composite, matte, checkerboard
AudioWhich sound the result keeps. Background video / Mix need a Background video connected.
Default: Foreground video · Choices: Foreground video, Background video, Mix both, Silent
Edge shrinkEat into the edge of the subject to lose fringing (pixels).
Default: 1 · Range 0 to 10
Edge featherSoften the edge of the matte (pixels).
Default: 1 · Range 0 to 10
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51

Presets: Remove black Remove white / sky Fire / smoke overlay Light leak Final: over background

Layout & Overlay

Stack 2×

Stack two videos vertically (input on top, input2 on bottom). Choose which video sets the length, and whose audio is kept.

Good for: A/B comparison videos · Performance or rehearsal footage with two simultaneous takes · Collage art that arranges multiple perspectives vertically

Settings

SettingWhat it does
Second VideoVideo placed on the bottom of the stack — connect a node or pick a file. Connect a second video to its extra dot, or add one with + Input video.
Bottom video fitThe bottom video always ends up as wide as the top one. This chooses how its height is handled.
Default: keep whole picture (scale to width) · Choices: keep whole picture (scale to width), fill and crop to the top video's height, stretch to the top video's size
Length set byLongest: the output runs as long as the longer video, and the shorter one holds its last frame. Top video / Bottom video: the output stops when that video stops, and the other is cut to match.
Default: Longest · Choices: Longest, Top video, Bottom video
AudioTop video or Bottom video keeps only that one's sound · Mix both plays the two together at equal level · Silent drops all sound.
Default: Top video · Choices: Top video, Bottom video, Mix both, Silent
Gap (px)Black gap between the two videos in pixels. Default: 0.
Default: 0 · Range 0 to 80

Tip: The cleaner the scaling match between the two videos, the sharper the final composite.

Layout & Overlay

Triptych

Arrange three videos in a symmetric hstack or vstack layout.

Settings

SettingWhat it does
Second VideoConnect a second video to its extra dot, or add one with + Input video.
Third VideoConnect a second video to its extra dot, or add one with + Input video.
LayoutDefault: hstack · Choices: hstack, vstack
Zoom 1Default: 1 · Range 0.1 to 5
Zoom 2Default: 1 · Range 0.1 to 5
Zoom 3Default: 1 · Range 0.1 to 5
Audio Mode1=video1 audio, 2=video2, 3=video3, 4=mix all, 5=mute, 6=external
Default: 1 · Choices: 1, 2, 3, 4, 5, 6
External AudioUsed when Audio Mode = 6 (external)
Volume 1Default: 1 · Range 0 to …
Volume 2Default: 1 · Range 0 to …
Volume 3Default: 1 · Range 0 to …
Layout & Overlay

Lagkage

Compose a base video with layered media defined by a JSON layout (images / videos / GIFs) using the lagkage compositor.

A JSON-driven multilayer video compositor that stacks, positions, resizes, mixes, sequences, and animates multiple media layers into a single rendered output. Lagkage is the most powerful structural tool in Videobeaux, allowing complex collage-based edits without touching an editor timeline.

Good for: Multi-video mosaics, walls, and grids · Dense collage compositions for live visual systems · Automated layout generation tools (e.g., Procedural 30-layout batches)

Settings

SettingWhat it does
Layout JSONThe layout file. Press ✎ to design it visually (drag images / GIFs / videos onto a stand-in for your video, or add text), or browse to an existing JSON.
Sequence DirectionOverride sequence_direction from JSON (forward|backward|random).
Choices: forward, backward, random
Audio ModeHow to build the audio track: base (default base video audio only), all (base + all JSON media audio mixed), json_only (only JSON media audio), external (use --audio-src), none (no audio).
Default: base · Choices: base, all, json_only, external, none
External audioUsed when Audio Mode = external.

Tip: Keep JSON clean and validated before rendering large batches.

Layout & Overlay

Layer Blend

Layer two videos with per-layer opacity and a blend mode — multiply, screen, color burn, difference, overlay and more. Choose whose audio to keep.

Settings

SettingWhat it does
Layer B (second video)The video layered on top of the main input (A) — connect a node or pick a file. Connect a second video to its extra dot, or add one with + Input video.
Opacity A (%)How much of the main input (A) shows.
Default: 75 · Range 0 to 100
Opacity B (%)How much of B (blended onto A with the mode) shows. In normal mode 75 / 25 = a 75% A + 25% B mix.
Default: 25 · Range 0 to 100
Blend modeBlend mode. Default: normal.
Default: normal · Choices: normal, multiply, screen, overlay, soft light, hard light, color dodge, color burn, darken, lighten, difference, exclusion, addition, subtract, divide, hard mix, vivid light, linear light, pin light, grain merge, grain extract, negation, average, phoenix, glow, heat, freeze, reflect
B sizecover = fill A's frame (crops), contain = fit inside (black bars), stretch = distort to fit.
Default: cover · Choices: cover, contain, stretch
LengthOutput length. Default: A (B loops if shorter).
Default: A (B loops if shorter) · Choices: A (B loops if shorter), shortest, longest
AudioWhich audio to keep: A, B, Mix A + B, or Silent. Default: A.
Default: A · Choices: A, B, Mix A + B, Silent
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51
Layout & Overlay

Picture-in-Picture

Picture-in-picture: a small inset video over the main one. Choose the corner (or exact position), size, border and opacity; use --swap to flip which video is full-screen; choose whose audio to keep.

Settings

SettingWhat it does
Inset videoThe small video shown inside the main one — connect a node or pick a file. Loops if shorter. Connect a second video to its extra dot, or add one with + Input video.
Size (% of width)Inset width as percent of the picture width. Default: 30.
Default: 30 · Range 5 to 100
PositionWhere the inset sits. Default: bottom right.
Default: bottom right · Choices: bottom right, bottom left, top right, top left, center, custom
Margin (%)Distance from the edge, as % of the width.
Default: 3 · Range 0 to 30
X (custom, %)Used when Position = custom.
Default: 50 · Range 0 to 100
Y (custom, %)Custom Y, percent of the way down (centre of the inset). Default: 50.
Default: 50 · Range 0 to 100
Border (px)Border width in pixels. Default: 4.
Default: 4 · Range 0 to 40
Border colorBorder color. Default: #FFFFFF.
Default: #FFFFFF
Opacity (%)Inset opacity, 0 to 100. Default: 100.
Default: 100 · Range 0 to 100
SwapMake the second video the full-screen picture and the main input the inset.
AudioA = main input, B = inset video.
Default: A · Choices: A, B, Mix A + B, Silent
Layout & Overlay

Quad Split

Split screen: up to four videos in a 2×2 grid, side by side, stacked, or one big picture plus three small. Empty inputs repeat the main video. Adjustable gap and color; choose whose audio to keep.

Settings

SettingWhat it does
Video 2Optional. Empty = repeats the main video. Connect a second video to its extra dot, or add one with + Input video.
Video 3Optional (layouts with 3+ cells). Connect a second video to its extra dot, or add one with + Input video.
Video 4Optional (layouts with 4 cells). Connect a second video to its extra dot, or add one with + Input video.
LayoutArrangement. Default: 2x2 grid.
Default: 2x2 grid · Choices: 2x2 grid, side by side (2), stacked (2), 1 big + 3 small
Gap (px)Gap between pictures in pixels. Default: 4.
Default: 4 · Range 0 to 60
Gap colorGap color. Default: #000000.
Default: #000000
FitHow each video fills its cell: cover crops, contain adds bars, stretch distorts.
Default: cover · Choices: cover, contain, stretch
AudioWhose audio to use. Default: Input 1.
Default: Input 1 · Choices: Input 1, Input 2, Input 3, Input 4, Mix all, Silent
Layout & Overlay

Video Wall

Video wall: the picture repeated in a grid of tiles — plain repeats, mirrored tiles, or a delay wall where every tile lags a bit more. Optional inverted checkerboard and gaps between tiles.

Settings

SettingWhat it does
ColumnsTiles across. Default: 3.
Default: 3 · Range 1 to 12
RowsTiles down. Default: 3.
Default: 3 · Range 1 to 12
Moderepeat = the same picture in every tile. mirror tiles = alternate tiles flipped so edges line up. delay wall = each tile is a little behind the previous one.
Default: repeat · Choices: repeat, mirror tiles, delay wall
Delay per tile (frames)For delay wall.
Default: 4 · Range 1 to 30
Invert every other tileInvert every other tile, checkerboard style.
Gap (px)Gap between tiles in pixels. Default: 0.
Default: 0 · Range 0 to 40
Gap colorGap color. Default: #000000.
Default: #000000
Layout & Overlay

Zoom

Zoom in by a percentage toward a point you choose. The frame keeps its size: the picture is cropped around the focus point and scaled back up.

Settings

SettingWhat it does
Zoom (%)100 = no zoom. 150 = 1.5× closer, 200 = twice as close.
Default: 150 · Range 100 to 400 · Good start: 110–250
Focus X (%)Horizontal point to zoom toward: 0 = left edge, 50 = centre, 100 = right edge.
Default: 50 · Range 0 to 100
Focus Y (%)Vertical point to zoom toward: 0 = top, 50 = middle, 100 = bottom.
Default: 50 · Range 0 to 100
Layout & Overlay

Crop Black Bars

Crop away black bars (top and bottom, or the sides) and keep the picture. The picture is then scaled to the nearest standard frame size (for example 1920×1080 or 1920×800), unless --keep_size is set.

Settings

SettingWhat it does
Black ThresholdHow dark a pixel must be to count as black (0–255). Raise it if the bars are not pure black.
Default: 24 · Range 4 to 64
Keep Cropped SizeKeep the picture at the size it was cropped to. Off = scale it to the nearest standard frame size.
Layout & Overlay

Rotate

Rotate the video: upside down, on its side, tilted at any angle, or slowly spinning. The output file really is rotated (its size follows).

Settings

SettingWhat it does
Angle (° clockwise)90 lays it on its side, 180 turns it upside down, 45 tilts it. Negative turns the other way.
Default: 180 · Range -180 to 180 · Good start: -180–180
FlipOptionally mirror the picture as well (applied before the rotation).
Default: None · Choices: None, Mirror left–right, Flip top–bottom
CornersOnly matters for angles that are not 90 / 180 / 270. Grow the canvas keeps everything; Zoom to fill enlarges the picture so there are no empty corners; Keep the frame size cuts the corners off.
Default: Grow the canvas (nothing cut off) · Choices: Grow the canvas (nothing cut off), Zoom to fill (no empty corners), Keep the frame size (corners cut off)
Empty-corner colorColor of empty corners. Default: #000000.
Default: #000000
Spin (° per second)Keeps turning while it plays, starting from the angle above. 0 = a still rotation.
Default: 0 · Range -360 to 360 · Good start: -90–90

Presets: Upside down On its side (right) On its side (left) Tilt 45° Slight tilt, filled Slow spin

Cut & Assemble

Trim

Extracts a single section of a video by timestamp — grab a clip from the middle, or trim off the start.

Settings

SettingWhat it does
StartSeconds (12.5) or HH:MM:SS(.ms). Default: start of video.
Default: 0
EndEnd timestamp. Mutually exclusive with Duration.
DurationLength to keep, from Start. Mutually exclusive with End.
Stream Copy (no re-encode)Much faster, but the cut snaps to the nearest keyframe rather than the exact timestamp.
Cut & Assemble

Qwikchop

Qwikchop (export-only): create EXACTLY N edits from each input video and place them in a folder. Optional leader trim for seamless heads. No merging.

Good for: Cutting commercials or interstitials from broadcast recordings · Building structured archives from long-form footage · Automatically splitting lecture or surveillance videos into events

Settings

SettingWhat it does
PiecesNumber of edits to create (exact count).
Range 1 to 200
RecurseIf input is a directory, include subdirectories.
Keep TempKeep temporary chunk files (debugging).
Trim Black FrontDetect and remove leading black at the start of each edit.
Black ScanSeconds to scan at each edit head for black (default: 0.25).
Default: 0.25 · Range 0.05 to 2
Black ThreshBlack luma threshold for blackdetect (0..1, default: 0.10).
Default: 0.1 · Range 0 to 1
Black PictMin fraction of black pixels to count as black (0..1, default: 0.10).
Default: 0.1 · Range 0 to 1
Edge Pad PreSeconds to shave from the very start of each edit (default: 0.03).
Default: 0.03 · Range 0 to 0.5
Edge Pad PostSeconds to shave from the end of each edit (default: 0.02).
Default: 0.02 · Range 0 to 0.5
Min EditMinimum allowed edit duration after trims (default: 0.20).
Default: 0.2 · Range 0.05 to 5
Cut & Assemble

Mince

Merge a folder of videos into one output in a chosen order.

A fast, precise segment extractor and reassembler that slices video into parts and recombines them with frame accuracy. Mince can cut, shuffle, normalize, re-encode, or stitch segments while preserving sync, enabling procedural editing, glitch slicing, and automated timeline reconstruction.

Good for: Procedural micro-edits for music videos · Rapid stutter-cut effects · Timeline reordering and random jump edits

Settings

SettingWhat it does
Order Modeforward/backward=filename order, lenfor/lenback=by duration, randn/randfib=random
Choices: forward, backward, lenfor, lenback, randn, randfib
EngineDefault: demuxer · Choices: demuxer, filter
Random SeedSeed for randn/randfib modes

Tip: For heavy glitch aesthetics, set high randomness with a fixed seed to remain reproducible.

Cut & Assemble

Concat

Concatenates two videos, first then second, back to back. Optionally crossfade between them, or insert a blank/silent gap. Good for adding a slate or title card before the main video.

Settings

SettingWhat it does
Second VideoPlays after the first video — connect a node or pick a file Connect a second video to its extra dot, or add one with + Input video.
Crossfade (s)Crossfade duration at the join. 0 = hard cut. Cannot be combined with Gap
Gap (s)Blank black + silent gap between the two clips. 0 = none. Cannot be combined with Crossfade
Cut & Assemble

Insert Clip

Inserts a second video into the master at a chosen timestamp, then picks up the master from where it left off — e.g. an intermission slate. Independent transition control at each boundary.

Settings

SettingWhat it does
Insert VideoVideo to insert into the master — connect a node or pick a file Connect a second video to its extra dot, or add one with + Input video.
Insert AtTimestamp in the master to insert at. Seconds (12.5) or HH:MM:SS(.ms)
Entry TransitionMaster → insert boundary
Default: cut · Choices: cut, crossfade, gap
Entry Duration (s)Crossfade or gap length at the entry boundary. Ignored for cut
Default: 0.5
Exit TransitionInsert → master boundary
Default: cut · Choices: cut, crossfade, gap
Exit Duration (s)Crossfade or gap length at the exit boundary. Ignored for cut
Default: 0.5
Cut & Assemble

Wipe Transitions

Combine two videos with a transitional wipe using ffmpeg's xfade filter.

Creates directional wipe transitions between clips using customizable timing, edge softness, and motion orientation

Settings

SettingWhat it does
Second VideoVideo that transitions in — connect a node or pick a file Connect a second video to its extra dot, or add one with + Input video.
Transition PresetDefault: fade · Choices: fade, wipeleft, wiperight, wipeup, wipedown, slideleft, slideright, slideup, slidedown, circlecrop, circleclose, circleopen, horizopen, horizclose, vertopen, vertclose, dissolve, pixelize, diagtl, diagtr, diagbl, diagbr, hlslice, hrslice, vuslice, vdslice, hblur, fadeblack, fadewhite, radial, smoothleft, smoothright, smoothup, smoothdown, rectcrop, distance, fadegrays, squeezeh, squeezev, zoomin
Transition Duration (s)Length of the wipe, taken from the end of the first video and the start of the second. Must fit inside both
Default: 1 · Range 0 to …
Cut & Assemble

Fade & Flash

Fade in and out from a color (black, white or any color) and add timed flashes. Flashes either decay smoothly or snap on and off. Optionally fades the audio with the picture.

Settings

SettingWhat it does
Fade in (s)Seconds to fade in from the color. Default: 1.
Default: 1 · Range 0 to 60
Fade out (s)Seconds to fade out to the color. Default: 1.
Default: 1 · Range 0 to 60
Fade colorFade color. Default: #000000.
Default: #000000
Fade audio tooFade the sound in and out along with the picture.
Flash at (seconds)Comma-separated times for flashes, e.g. 1.5, 4, 6.25. Leave empty for none.
Flash length (s)Flash length in seconds. Default: 0.25.
Default: 0.25 · Range 0.04 to 5
Flash colorFlash color. Default: #FFFFFF.
Default: #FFFFFF
Flash styleFlash style. Default: decay.
Default: decay (fades away) · Choices: decay (fades away), hard (on/off)
Cut & Assemble

Add Audio

Put your own music or audio under a video — replace the sound, mix it in, or duck it under the original voice.

Settings

SettingWhat it does
Music / audio fileAny audio file (mp3, wav, m4a, flac…) — or a video file, to use its sound.
How to combineReplace = only your audio. Mix = your audio plays together with the original. Duck = like Mix, but your audio automatically dips while the original is loud (speech).
Default: Replace the original sound · Choices: Replace the original sound, Mix under the original, Duck under the original (voice stays clear)
Music volume (%)Soundtrack volume in percent. Default: 100.
Default: 100 · Range 0 to 200 · Good start: 40–100
Original volume (%)Only used when mixing or ducking.
Default: 100 · Range 0 to 200 · Good start: 60–100
Start music at (s)Delay before the music begins.
Default: 0 · Range 0 to 3600
Skip into the music (s)Start playing the audio file from this point.
Default: 0 · Range 0 to 3600
Fade in (s)Fade-in seconds. Default: 0.
Default: 0 · Range 0 to 30
Fade out (s)Fade-out seconds at the end of the video. Default: 0.
Default: 0 · Range 0 to 30 · Good start: 1–4
If the music is shorterLoop the music to fill the video, or play it once.
Default: Loop to fill the video · Choices: Loop to fill the video, Play once
Duck strengthHow hard the music dips under the original sound (Duck mode).
Default: 8 · Range 1 to 20 · Good start: 4–10

Presets: Replace Music bed (quiet) Voice-over friendly (duck)

Speech & Captions

Audio Duck

Audio Duck: lower the music under the narration. The clip's own sound is the voice; the connected music drops whenever the voice is speaking, then comes back up. The picture is copied untouched.

Settings

SettingWhat it does
MusicConnect the music (any video or audio file). It loops if it is shorter than the clip. Connect a second video to its extra dot, or add one with + Input video.
Speech level (dB)The level the voice must reach before the music ducks. Quiet narration needs a lower number.
Default: -35 · Range -60 to -10 · Good start: -45–-20
Duck amount (ratio)How hard the music is pulled down under speech. Higher = the music drops further.
Default: 6 · Range 1 to 20 · Good start: 3–10
Release (ms)How long the music takes to come back up after the voice stops.
Default: 400 · Range 50 to 2000 · Good start: 200–800
Music level (dB)Level of the music before ducking. Negative = quieter.
Default: -6 · Range -30 to 6 · Good start: -18–0
Speech & Captions

Subtitle Burn

Subtitle Burn: burn an .srt subtitle file into the picture with the font, size, colours, outline and position you choose. The sound is kept.

Settings

SettingWhat it does
Subtitle file (.srt)The subtitle file to burn in. Times are read from the file.
FontA font name installed on this computer, for example Arial.
Default: Arial
Font sizeSubtitle size in pixels, as the video is drawn.
Default: 24 · Range 8 to 120 · Good start: 16–60
Text colorColour of the subtitle text.
Default: #FFFFFF
Outline colorColour of the outline around the text.
Default: #000000
Outline widthThickness of the outline, in pixels. 0 = none.
Default: 2 · Range 0 to 6 · Good start: 0–4
PositionWhere the subtitles sit on the picture.
Default: bottom · Choices: bottom, middle, top
Margin from the edge (px)Distance from the bottom or top edge. Ignored for middle.
Default: 30 · Range 0 to 300 · Good start: 0–120
Speech & Captions

Loudness Normalize

Loudness Normalize: bring the sound to a target loudness (EBU R128, in LUFS) with a true-peak ceiling, so quiet and loud clips come out at the same level. The picture is copied untouched.

Settings

SettingWhat it does
Target loudness (LUFS)-16 is the usual level for online video and podcasts; -23 is the broadcast standard.
Default: -16 · Range -30 to -5 · Good start: -24–-12
Peak ceiling (dBTP)The loudest peak allowed after normalising. -1.5 leaves headroom so the sound cannot clip.
Default: -1.5 · Range -9 to 0 · Good start: -3–-1
Loudness range (LU)How much the loudness may vary within the clip. Smaller is more even, but can sound pumped.
Default: 11 · Range 1 to 20 · Good start: 5–15
Speech & Captions

Audio Denoise

Audio Denoise: take hiss and hum out of speech. A high-pass removes low rumble, then an FFT denoiser lowers the background noise. The picture is copied untouched.

Settings

SettingWhat it does
Reduce noise by (dB)How far the noise is pushed down. More cleans more, but a voice can start to sound watery.
Default: 12 · Range 0 to 40 · Good start: 6–20
Noise floor (dB)The level of the background noise. Set it near the hiss you hear in the quiet parts.
Default: -25 · Range -80 to -20 · Good start: -50–-25
Cut rumble below (Hz, 0 = off)Removes low rumble and hum under the voice. 0 leaves the low end alone.
Default: 80 · Range 0 to 300 · Good start: 40–150
Speech & Captions

Transcraibe

AI speech-to-text transcription (Vosk).

Settings

SettingWhat it does
Vosk ModelPath to the extracted Vosk model directory
Also Write .txtWrite a plain-text version of the transcript alongside the JSON
Overwrite Existing
Speech & Captions

Captburn

Burns the clip's own speech onto the picture as kinetic-typography captions: the spoken word pops larger and changes colour, in any font you pick. Uses the clip's transcript if there is one; otherwise transcribes it locally (Vosk) first.

Burns subtitles, captions, or transcript text directly into the video with precise styling, timing, and layout control

Settings

SettingWhat it does
Vosk ModelUsed to transcribe the clip first. Only needed when the clip has no transcript (<clip>.json) yet.
FontPick a font — each one is shown in its own typeface. The bundled ones work the same on every computer.
Default: Arial
Font SizeRequested size in px — shrunk automatically so every line fits the frame.
Default: 76 · Range 12 to 200 · Good start: 48–110
Text ColorWords that are not being spoken right now.
Default: #FFFFFF
Highlight ColorThe word being spoken.
Default: #FFE128
Outline ColorCaption text outline color. Default: #000000 (black).
Default: #000000
Outline ThicknessPixels.
Default: 8 · Range 0 to 24 · Good start: 4–12
Words per CaptionHow many words are on screen together. Short bursts read best.
Default: 4 · Range 1 to 10 · Good start: 2–5
Highlight PopHow much bigger the spoken word is drawn. 1 = same size.
Default: 1.18 · Range 1 to 2 · Good start: 1.1–1.4
Keep CapitalisationLeave the words as transcribed instead of ALL CAPS.
Vertical PositionWhere the captions sit: 0 = top of the picture, 1 = bottom.
Default: 0.67 · Range 0.05 to 0.95 · Good start: 0.5–0.85
Max WidthWidest a caption may be, as a share of the picture width.
Default: 0.86 · Range 0.3 to 1
Max LinesThe font shrinks until the caption fits in this many lines.
Default: 2 · Range 1 to 4
Speech & Captions

Auto Narrate

Adds AI-narrated captions to a video — type your own script (offline, default) or draft one from a topic via an optional local Ollama model. Synthesizes speech (kokoro-tts), transcribes it locally (Vosk), mixes it over the original audio, and burns kinetic-typography captions (the spoken word pops larger, per-frame rendered — see Kinetic Captions).

Settings

SettingWhat it does
ScriptYour own narration text, used verbatim. The default, no-setup path — works fully offline. Add <<p:2>> for two seconds of silence at that point, or <<p:.5>> for half a second.
Topic (optional)Alternative to Script: draft narration from a topic via the optional Ollama Model below.
Ollama Model (optional)Only used with Topic — pick a locally-pulled Ollama model. Leave blank if using Script.
Target Duration (s)Only steers Ollama generation when Topic is used
Default: 30
Voicekokoro-tts voice — listed live from your local kokoro-tts install
Default: am_adam
Voice Speedkokoro-tts speech speed multiplier
Default: 1
Narration Starts At (s)Seconds of silence before the narration starts, so it begins that long after the video does. The video itself is not shifted
Default: 0 · Range 0 to …
Vosk ModelUsed to transcribe the synthesized narration
Original Audio VolumeBackground level of the original audio under the narration, 0-1
Default: 0.2
Duck the video audio under the narrationLowers the video's own sound whenever the narration speaks, so the voice stands out. Uses Duck Ratio.
Duck RatioHow hard the video's sound is pulled down under the narration (1 = barely, 20 = almost silent). Only used with Duck.
Default: 6 · Range 1 to 20
FontCaption font name (or a path to a .ttf/.otf file)
Default: Arial
Font SizeRequested caption font size in px — auto-shrunk as needed to fit the frame
Default: 76
Text ColorColor of words that are not currently being spoken
Default: #FFFFFF
Highlight ColorColor of the word currently being spoken
Default: #FFE128
Outline ColorCaption text outline color
Default: #000000
Speech & Captions

Grep Supercut

Searches a transcript for a word, phrase, or regex and cuts together every match — the classic "videogrep": find every time someone says X.

Settings

SettingWhat it does
Vosk ModelUsed to transcribe the video first, if no transcript for it exists yet
Search QueryWord, phrase, or regex to search for. Comma-separated for multiple queries.
Search Typesentence: whole matching lines. fragment: exact matched phrase. mash: random word-mashup.
Default: sentence · Choices: sentence, fragment, mash
Whole Word OnlyMatch "cut" as a whole word only — turn off to search as a raw regex/substring (e.g. "cut.*ing"). No effect in mash mode.
Default: On
Padding (s)Seconds added before/after each matched clip
Default: 0
Max Clips0 = unlimited
Default: 0
Randomize OrderShuffle the matched clips before assembling the supercut
Speech & Captions

Qwikchop Deluxe

Content-aware highlight extraction: scores candidate excerpts from the transcript (local centrality + audio energy + keyword hooks — no network) and exports the best ones, each as its own file.

Settings

SettingWhat it does
Trans JsonTranscript JSON to use (default: <input>.json).
Vosk ModelUsed to transcribe the video first, if no transcript for it exists yet.
CountNumber of excerpts to export. Default: 5.
Default: 5 · Range 1 to 50
Target DurationIf set (>0), keep adding top-scoring excerpts (skipping any that would overflow) until this total duration is reached, instead of a fixed --count. --count still caps the count if both are set.
Default: 0 · Range 0 to 600
Min TimeShortest allowed excerpt, in seconds. Default: 3.
Default: 3 · Range 0.5 to 60
Max TimeLongest allowed excerpt, in seconds. Default: 20.
Default: 20 · Range 1 to 120
Min GapMinimum seconds of separation between selected excerpts, so picks don't cluster on the same moment. Default: 2.
Default: 2 · Range 0 to 60
PaddingSeconds of padding added before and after each exported excerpt. Default: 0.
Default: 0 · Range 0 to 5
Hook WordsComma-separated words that nudge an excerpt's score up when present. Default: a generic hook-word list — tune it for your content.
Default: secret,never,actually,literally,wow,insane,worst,best,amazing,shocking,unbelievable,crazy,honestly,wild
Weight CentralityWeight of the local TextRank-style centrality score, 0-1ish. Default: 0.5.
Default: 0.5 · Range 0 to 1
Weight EnergyWeight of the audio-energy score, 0-1ish. Default: 0.3.
Default: 0.3 · Range 0 to 1
Weight HooksWeight of the keyword-hook score, 0-1ish. Default: 0.2.
Default: 0.2 · Range 0 to 1
Ollama Model (optional)Optional local re-ranking of the shortlisted excerpts. Leave blank to use the local scoring only.
Llm CandidatesHow many top locally-scored candidates to hand to --llm_model for re-ranking. Default: 15.
Default: 15 · Range 3 to 60
VcodecDefault: libx264
CrfDefault: 18 · Range 0 to 51
PresetDefault: veryfast
Speech & Captions

Silence Xtraction

Remove speech segments, keeping what's left (not exactly silence, but no discernable words).

Settings

SettingWhat it does
Vosk ModelUsed to transcribe the video first, if no transcript for it exists yet
Min Silence Duration (s)Minimum duration of a silence to consider
Max Silence Duration (s)Maximum duration of a silence to consider
End Adjuster (s)Offset to shorten the silence from the end — closer to 0 includes more non-word sounds
Speech & Captions

Word Xtraction VAD

Remove recognized speech (Vosk) and optionally local Silero VAD detections, keeping applause, laughter, music, noise, and other non-speech audio.

Settings

SettingWhat it does
Vosk ModelUsed to transcribe the video first, if no transcript for it exists yet
Min Region Duration (s)Minimum duration of a kept region
Max Region Duration (s)Maximum duration of a kept region. With Keep Non-Speech, use 0 or negative for no maximum
End Adjuster (s)Legacy-mode offset that shortens a gap from the end. Ignored with Keep Non-Speech — use Speech Padding instead
Keep Non-SpeechKeep everything outside recognized speech, including material before the first word and after the last — the preferred mode for retaining applause, laughter, music, ambience, and noise
Default: On
Max Word Duration (s)In Keep Non-Speech mode, longer Vosk word timestamps are treated as suspicious and clamped
Default: 1.5
Speech Padding (s)Padding added before/after each recognized word in Keep Non-Speech mode, to avoid clipped syllables
Default: 0.08
Merge Gap (s)Merge neighboring kept regions separated by this many seconds or less
Default: 0.2
Use Silero VADLocally detect speech even when Vosk cannot recognize the words (requires the silero-vad package; runs entirely locally)
Default: Off
VAD ThresholdSilero speech probability threshold — lower is more aggressive
Default: 0.45
VAD Min Speech (ms)Default: 150
VAD Min Silence (ms)Default: 120
VAD Padding (ms)Padding around each VAD speech region
Default: 120
VAD Keep Short (s)Preserve brief VAD-only vocal texture (ums, uhs, breaths) at or below this duration; recognized Vosk words are still removed regardless. 0 disables
Default: 0.4
Speech & Captions

Ngrams

Lists the most common word sequences spoken in a transcript — useful for finding good search terms before running Grep Supercut.

Settings

SettingWhat it does
Vosk ModelUsed to transcribe the video first, if no transcript for it exists yet
N-gram Size1 for single words, 2 for word pairs, etc.
Default: 2
Top ResultsHow many ranked results to keep
Default: 30
Media Tools

Sharpen

Sharpen: crisp up soft footage. unsharp lifts edges over a chosen detail size; cas adds detail where the picture is soft and leaves crisp edges alone. Colour is not changed.

Settings

SettingWhat it does
Methodunsharp = classic edge sharpening. cas = contrast-adaptive: sharpens the soft areas and leaves already-crisp edges alone, so it adds less halo.
Default: unsharp · Choices: unsharp, cas
Amount0 = none. Around 0.5 is subtle; above 1.5 halos show around edges.
Default: 0.8 · Range 0 to 3 · Good start: 0.3–1.5
Detail size (px)Size of the detail being sharpened (unsharp only). Small = fine edges, large = broader ones. Odd numbers.
Default: 5 · Range 3 to 13 · Good start: 3–7
Media Tools

Deinterlace

Turns interlaced footage (the comb lines on camcorder and broadcast video) into clean frames, with yadif or bwdif.

Settings

SettingWhat it does
Methodyadif is quick and reliable. bwdif keeps a little more fine detail, but is slower.
Default: yadif · Choices: yadif, bwdif
One frame per field (double rate)Doubles the frame rate so every field becomes a frame: the smoothest motion. Off = one frame per frame.
Media Tools

Denoise Video

Denoise: soften the grain and speckle of noisy footage (low light, high ISO) while keeping the edges, by smoothing within each frame and across frames.

Settings

SettingWhat it does
Strengthlight keeps most of the detail. strong smooths a lot and can look painterly.
Default: medium · Choices: light, medium, strong
Media Tools

Lens Correct

Lens Correct: straighten the bend of a wide-angle or action-camera lens. Negative k1 removes barrel distortion (edges bowing out); positive k1 removes pincushion (edges pinched in). k2 is a fine tune.

Settings

SettingWhat it does
Correction k1The main correction. Try negative values for a barrel bulge (wide lenses, action cameras); positive values for pinched edges.
Default: -0.2 · Range -0.5 to 0.5 · Good start: -0.4–0.1
Correction k2A fine correction for the far edges. Set it after k1.
Default: 0 · Range -0.5 to 0.5 · Good start: -0.2–0.2
Media Tools

Convert Format

Black-box FFmpeg converter: any format in → mp4/mov/avi/mkv/webm out (container chosen by the -o extension).

Rewraps or converts media streams while copying or re-encoding video/audio. Useful for repairing containers, adjusting codecs, changing formats, resolving sync issues, or preparing files for specific platforms or pipelines.

Good for: Generating GIFs, PNG/JPEG sequences, or audio-only MP3/AAC/FLAC assets

Settings

SettingWhat it does
ModeSimple: pick a target format, size and quality and everything else is chosen for you. Advanced: set codecs, bitrates and filters yourself.
Default: Simple (pick a format) · Choices: Simple (pick a format), Advanced (every setting)
Convert toSet the Output node's Container format to match (MP4, MOV, WEBM, MKV, GIF, MP3, WAV or M4A).
Default: MP4 · H.264 — plays everywhere · Choices: MP4 · H.264 — plays everywhere, MP4 · HEVC (H.265) — smaller files, MOV · ProRes 422 — for editing, WebM · VP9 — for the web, MKV · H.264, GIF — short looping picture, MP3 — audio only, WAV — audio only (uncompressed), M4A — audio only (AAC)
SizeSimple mode: output size.
Default: Original size · Choices: Original size, 4K wide (3840), 1080p wide (1920), 720p wide (1280), 480p wide (854), Square 1080×1080 (crops), Vertical 1080×1920 (crops), Custom width × height
Custom width (px)Simple mode, custom size: width in pixels.
Default: 0 · Range 16 to 7680
Custom height (px)Use -2 style "keep aspect" by leaving 0.
Default: 0 · Range 16 to 4320
QualityHigher quality = larger file. Balanced is right for almost everything.
Default: Balanced · Choices: High (large file), Balanced, Small file
Container hintForce the container (mp4, mov, webm, matroska…). Normally taken from the Output node.
Quick PresetOverrides the codec/quality settings below. Leave blank to use them instead. Must match the Output node's format.
Choices: mp4_h264, mp4_hevc, mp4_av1, webm_vp9, webm_av1, prores_422, prores_4444, dnxhr_hq, lossless_ffv1, avi_mjpeg_fast, avi_mpeg4_fast
Video CodecVideo codec (e.g., libx264, libx265, libaom-av1, prores_ks, dnxhd, mpeg2video, mpeg4, mjpeg)
Default: libx264 · Choices: libx264, libx265, libvpx-vp9, libsvtav1, prores_ks, mpeg4, mjpeg, ffv1, dnxhd
Audio CodecAudio codec (e.g., aac, libopus, libmp3lame, mp3, pcm_s16le)
Default: aac · Choices: aac, libmp3lame, libopus, flac, pcm_s16le, pcm_s24le
Quality (CRF)Lower = higher quality. Tuned for the default libx264 codec — x265/VP9/AV1 typically want higher values (~28-35).
Default: 23 · Range 0 to 51
Video Bitratee.g. 5M. Overrides CRF-based quality if set.
Video Max RateVideo maxrate
VBV Buffer SizeVBV buffer size
Encoder Presetx264/x265 speed-vs-quality preset. Leave blank for other codecs.
Default: medium · Choices: ultrafast, superfast, veryfast, faster, fast, medium, slow, slower, veryslow
Codec Profilex264/x265 profile restriction. Leave blank to let the encoder choose.
Choices: baseline, main, high, high10, high422, high444
Codec LevelLeave blank to let the encoder choose.
Choices: 3.0, 3.1, 4.0, 4.1, 4.2, 5.0, 5.1, 5.2
Pixel FormatPixel format (e.g., yuv420p, yuv422p10le, yuva444p10le)
Default: yuv420p · Choices: yuv420p, yuv422p, yuv444p, yuv420p10le, yuv422p10le, yuva444p10le
GOP SizeKeyframe interval in frames.
Range 1 to 600
Frame RateLeave blank to keep the source frame rate.
Choices: 23.976, 24, 25, 29.97, 30, 50, 59.94, 60
Extra ffmpeg filters (optional)A raw ffmpeg video filter chain, e.g. scale=1280:-2,eq=contrast=1.1 — leave empty if unsure.
Video Tag (fourcc)e.g. hvc1 for HEVC in MP4.
Audio BitrateAudio bitrate (e.g., 192k)
Default: 192k
Audio ChannelsAudio channels
Default: 2 · Choices: 1, 2, 6, 8
Audio Sample RateAudio sample rate (e.g., 48000)
Default: 48000 · Choices: 22050, 44100, 48000, 96000
Stream Copy (no re-encode)Skip re-encoding when the source already fits (fast, lossless). Ignores the settings above.

Presets: Web · small MP4 Share · 720p MP4 Editing · ProRes Audio · MP3 GIF · 480

Tip: Use the ProRes and DNxHR profiles for grading/VFX workflows where multi-gen quality matters.

Media Tools

Convert Dims

Converts an input video to a specified preset dimension. Choose how to handle aspect ratio: pad, fit, fill-crop, or stretch. Use --portrait-full to force a 9:16 full-frame vertical output.

, 1080p, instagram_reels, tiktok_video) and applies a configurable aspect-ratio handling mode (pad, fit, fill, fill_crop, stretch). Supports portrait overrides, crop/pad anchoring, custom pad color, and scaling kernel selection.

Good for: Generating social-media-optimized versions (Reels, Shorts, Stories, TikTok) · Quickly standardizing deliverables to 1080p, 4K, or other common presets

Settings

SettingWhat it does
Output FormatFormat for output file extension (e.g., mp4, mov). Optional: defaults to the output file's own extension, or mp4.
Size presetPick a common size and the width and height fields fill in. Pick Custom to type your own.
Default: 1080p · Choices: sd, 720hd, 1080hd, widescreen, portrait1080, 480p, 576p, 720p, 1080p, 1440p, 4k, 8k, vga, qvga, wvga, svga, xga, wxga, sxga, uxga, wuxga, qwxga, qhd, wqxga, 5k, portrait720, portrait4k, square1080, square720, cinema4k, ultrawide1080, ultrawide1440, instagram_feed, instagram_reels, instagram_stories, tiktok_video, youtube_standard, youtube_shorts, facebook_feed, facebook_stories, twitter_video, twitter_square, linkedin_video, linkedin_square, snapchat_video, pinterest_video, pinterest_square, custom
Width (px)Output width. Used with Height; overrides the preset.
Default: 1920 · Range 16 to 8192
Height (px)Output height. Used with Width; overrides the preset.
Default: 1080 · Range 16 to 8192
Modepad: keep AR + borders; fit: keep AR, no pad; fill/fill_crop: cover + center-crop; stretch: ignore AR.
Choices: pad, fit, fill, fill_crop, stretch
Translate(Deprecated) yes=stretch, no=pad if --mode not set.
Default: no · Choices: yes, no
AnchorWhere to bias the crop or padding.
Default: center · Choices: center, top, bottom, left, right, top_left, top_right, bottom_left, bottom_right
Pad ColorPadding color for --mode pad. Hex like #000000, #FFFFFF.
Default: #000000
Scale FlagsScaler kernel for the scale filter.
Default: lanczos · Choices: lanczos, bicubic, bilinear, neighbor
Portrait FullForce 9:16 portrait output and fill (cover + crop).
Media Tools

Color & Pattern

Makes a video of a solid colour, gradient, checkerboard, colour bars, test pattern or noise, for as many seconds as you set. No input video needed: connect it straight to Output.

Settings

SettingWhat it does
PatternSolid colour uses the colour below. Gradient and Checkerboard use it too. The bars, test pattern, RGB ramp and noise have their own colours.
Default: Solid colour · Choices: Solid colour, Gradient (colour to black), Checkerboard, SMPTE colour bars, SMPTE HD colour bars, Test pattern, RGB ramp, Noise
ColourThe colour for Solid colour, Gradient and Checkerboard.
Default: #FF0000
InputMake a picture: the picture alone, for the length below. Over the input: the picture on top of the video connected to this node. Behind the input: the picture underneath it, showing through wherever the video is transparent. With an input, the size and length come from the video.
Default: Make a picture · Choices: Make a picture, Behind the input, Over the input
OpacityHow strong the picture is over or behind the input video, 0 to 1. Only used with an input.
Default: 1 · Range 0 to 1
Length (seconds)How long the picture runs when no input is connected.
Default: 5 · Range 0.1 to 600
Frame rateFrame rate when no input is connected. With an input, its frame rate is used.
Default: 30 · Range 1 to 120
Size presetPick a common size and the width and height fill in. Pick Custom to type your own. Not used when an input is connected.
Default: 1080p · Choices: sd, 720hd, 1080hd, widescreen, portrait1080, 480p, 576p, 720p, 1080p, 1440p, 4k, 8k, vga, qvga, wvga, svga, xga, wxga, sxga, uxga, wuxga, qwxga, qhd, wqxga, 5k, portrait720, portrait4k, square1080, square720, cinema4k, ultrawide1080, ultrawide1440, instagram_feed, instagram_reels, instagram_stories, tiktok_video, youtube_standard, youtube_shorts, facebook_feed, facebook_stories, twitter_video, twitter_square, linkedin_video, linkedin_square, snapchat_video, pinterest_video, pinterest_square, custom
Width (px)Output width. Used with Height; overrides the preset. Not used when an input is connected.
Default: 1920 · Range 16 to 8192
Height (px)Output height. Used with Width; overrides the preset. Not used when an input is connected.
Default: 1080 · Range 16 to 8192

Presets: Colour bars HD colour bars Test pattern Black White Red Green Blue Yellow Red / black checks

Media Tools

Resize

Akin to repainting the same image while smudged with alcohol.

Good for: Creating platform-targeted versions: 1080×1920 (reels/stories/shorts) · 1080×1080 (square grid formats) · 1920×1080 (landscape)

Settings

SettingWhat it does
Width (px)Target width in pixels
Height (px)Target height in pixels

Tip: For crisp results, resize before applying heavy effects and grading.

Media Tools

Shortify

Converts a video to vertical Shorts/Reels format with a blurred, enlarged background fill instead of black bars.

Settings

SettingWhat it does
FormatIgnored if Custom Width/Height are both set.
Default: shorts_1080x1920 · Choices: shorts_1080x1920, portrait_1080x1350, square_1080x1080
Custom WidthOverrides Format when set together with Custom Height.
Custom Height
Blur Strength0 = sharp enlarged background, no blur.
Default: 20
Media Tools

Tonemap HDR→SDR

Tonemap HDR content to SDR.

Converts HDR footage (PQ/HLG) to SDR using tunable tonemapping curves, preserving highlight detail and color accuracy.

Good for: Creating SDR masters from HDR originals · Preparing high-quality web deliverables (YouTube, Vimeo, TikTok) · Producing SDR screeners or festival submissions that disallow HDR

Settings

SettingWhat it does
AlgoTonemap operator. Default: hable
Default: hable · Choices: hable, mobius, reinhard, clip
DesatDesaturate highlights during tonemap [0.0–1.0]. Default: 0.0
Default: 0 · Range 0 to 1
PeakPeak brightness of the HDR source in nits (1000 is typical, 4000 for some masters). Default: 1000
Default: 1000 · Range 100 to 10000
DitherDither mode applied in zscale prior to format(). Default: error_diffusion
Default: error_diffusion · Choices: none, ordered, random, error_diffusion
Pix FmtOutput pixel format. Common picks: yuv420p, yuv422p10le. Default: yuv420p
Default: yuv420p
X264 Presetlibx264 preset (if re-encoding). Default: medium
Default: medium
CrfCRF when encoding with libx264. Default: 18
Default: 18 · Range 0 to 51
Copy AudioCopy audio stream instead of re-encoding.

Tip: Use yuv422p10le when creating grading intermediates.

Media Tools

Extract Frames

Extract frames from video as images.

Extracts individual frames from a video file and saves them as image files (typically PNG). Ideal for analysis, archival, animation workflows, visual debugging, and creating frame-based artwork.

Good for: Creating frame sequences for animation, compositing, or painting · Building datasets for computer vision or machine learning · Inspecting motion continuity, exposure behavior, or compression artifacts

Settings

SettingWhat it does
Frame RateThe number of PNGs extracted per second of video

Tip: Use PNG for archival quality; JPG for lightweight previews.

Media Tools

Extract Sound

Extracts the audio as a .wav file from the --input video file.

Extracts the audio track from a video file and saves it as a standalone audio file, preserving quality and format. Useful for sound design, transcription workflows, remixing, archival storage, or preparing audio-only deliverables.

Good for: Extracting dialogue or interviews for transcription · Creating podcast audio from filmed sessions · Pulling music or SFX layers from video renders

Settings

SettingWhat it does
Sample rate (Hz)44100 = CD quality (default) · 48000 = video standard · 22050 = small voice-only files.
Default: 44100 · Choices: 22050, 32000, 44100, 48000
ChannelsMono mixes both channels into one (half the size).
Default: Stereo · Choices: Stereo, Mono

Tip: If audio is distorted or clipped, normalize or limit in a separate audio-processing step.

Media Tools

Thumbs

Generate thumbnails and/or a tiled contact sheet from a video. Supports interval or scene-based selection, timestamps, custom fonts, and layout styling.

Generates thumbnails or contact sheets by sampling frames at chosen intervals for previews, galleries, or QC review.

Good for: QC review workflows to quickly scan for artifacts · Preview libraries for editors, designers, or curators · Thumbnail generation for apps, web galleries, or dataset creation

Settings

SettingWhat it does
FpsFrames per second to sample (e.g., 0.5 = one every 2s).
Default: 0.5 · Range 0.1 to 10
SceneUse scene-based selection instead of fixed intervals.
Scene ThresholdScene detection sensitivity (lower = more cuts).
Default: 0.4 · Range 0 to 1
TileContact sheet grid 'COLUMNSxROWS' (e.g., 6x4).
Default: 6x4
ScalePer-thumb scale (e.g., 320:-1).
Default: 320:-1
TimestampsOverlay timestamps on thumbnails.
LabelAdd a footer label with filename.
FontfileCustom font path for drawtext.
BgBackground color ('black', '#111111', or '0x111111').
Default: #000000
MarginOuter margin (pixels).
Default: 12 · Range 0 to 100
PaddingPadding between tiles (pixels).
Default: 6 · Range 0 to 100
OutdirDirectory to export frame sequence.
Image FormatOutput format if --outputfile has no extension.
Choices: jpg, png
Jpeg QualityJPEG quality (2=high, 3=good, 5=ok).
Default: 3 · Range 1 to 31

Tip: For clean contact sheets, use PNG for lossless quality.

Media Tools

Subs Convert

List, extract, and convert subtitle tracks. Works with container-embedded subtitles or standalone .srt/.ass/.vtt files.

, SRT, VTT, ASS) while preserving timing, text, and (where possible) style metadata.

Settings

SettingWhat it does
ListList subtitle streams in the input video and exit.
IndexesComma-separated list of subtitle stream indexes to extract (e.g., '0,2').
LangsComma-separated list of language codes to extract (e.g., 'eng,spa').
AllExtract all subtitle streams.
Forced OnlyOnly include streams with 'forced' disposition.
Exclude HiExclude streams with 'hearing_impaired' disposition.
FormatTarget subtitle format for output. Required for standalone subtitle conversion; optional for video mode.
Choices: srt, vtt, ass
OutdirDirectory for multiple extracted subtitle files.
Time ShiftApply time shift in seconds (can be negative).
Default: 0 · Range -60 to 60
Media Tools

Media Info

Concise media report: dimensions, duration, fps, codecs, bitrate, size and audio details. Prints a readable summary and the JSON to the log (copy it from the console) and writes the JSON file.

Settings

SettingWhat it does
No JsonOnly print the summary; don't write a JSON file.
Media Tools

Meta Extraction

Extract extensive metadata to JSON, with optional analysis passes (frame sampling, blackdetect, EBU R128 loudness).

Extracts detailed metadata—such as codecs, bitrates, dimensions, color information, loudness, black-frame metrics, and stream structure—from any media file.

Good for: Preflight analysis before encoding or delivery · QC checks for broadcast, festival, or archival compliance · Automated pipelines needing structural insight into media files

Settings

SettingWhat it does
Sample FramesEnable frame sampling via ffprobe (pict_type, keyframe, WxH).
Sample StrideSampling stride in seconds (default 0.5).
Default: 0.5 · Range 0.1 to 10
Sample LimitMax sampled frames to keep (default 200).
Default: 200 · Range 10 to 2000
BlackdetectEnable black frame detection summary.
Black Pic Thblackdetect picture threshold (default 0.10).
Default: 0.1 · Range 0 to 1
Black Dur MinMinimum black duration in seconds (default 0.10).
Default: 0.1 · Range 0 to 5
LoudnessRun EBU R128 loudness analysis (integrated, LRA, true peak).
Media Tools

Hash Fingerprint

Compute file hashes (md5/sha1/sha256), optional FFmpeg stream hash, per-frame checksums, and optional perceptual hashes over sampled frames. Works on a single input or a directory.

Creates unique fingerprints of video files using checksums, perceptual hashing, or frame-level hashing. Useful for deduplication, archival identification, similarity matching, and database cataloging.

Good for: Archival fingerprinting for media libraries · Deduplication of large video collections · Detecting alternate encodes of the same content

Settings

SettingWhat it does
RecursiveIf input is a directory, scan recursively.
ExtsFile extensions to include when scanning a directory (case-insensitive).
File HashesFile-level hashes to compute (streamed, no load into RAM).
Choices: md5, sha1, sha256
Stream HashUse FFmpeg -f hash to hash the primary video stream (fast, codec-level).
Default: none · Choices: none, md5, sha256
Framemd5Emit per-frame checksums via FFmpeg -f framemd5 (verbose).
PhashCompute perceptual hashes (average hash) over sampled frames. Requires Pillow.
Phash FpsApprox frames-per-second to sample for perceptual hashing (0.5 = one frame every 2s).
Default: 0.5 · Range 0.1 to 10
Phash SizeaHash size NxN (default 8 -> 64-bit hash).
Default: 8 · Range 4 to 32
Stream KindWhich primary stream to hash for --stream-hash/--framemd5.
Default: video · Choices: video, audio
Media Tools

Stabilize

Video stabilization: tracks camera motion with feature points, smooths the camera path, and re-renders each frame along the smoothed path. Zooms in slightly to hide the edges.

Settings

SettingWhat it does
SmoothnessSmoothing window in frames — larger = steadier, floatier camera. Default: 30.
Default: 30 · Range 1 to 120
ZoomZoom to hide moving borders (1.0 = none, shows edges). Default: 1.06.
Default: 1.06 · Range 1 to 1.5
Analysis WidthWidth the motion is analyzed at (smaller = faster). Default: 480.
Default: 480 · Range 160 to 1280
Crfx264 quality. Default: 18.
Default: 18 · Range 0 to 51