# ComfyTV

## Docs

- [Bridge nodes — connecting other ComfyUI plugins](https://comfytv.org/guides/bridges.md)
- [Composing & arranging](https://comfytv.org/guides/compose.md)
- [Custom workflows](https://comfytv.org/guides/custom-workflows.md)
- [Generating content](https://comfytv.org/guides/generate.md)
- [Getting started](https://comfytv.org/guides/getting-started.md)
- [Image tools](https://comfytv.org/guides/image-tools.md)
- [Making Music in ComfyTV/ComfyUI](https://comfytv.org/guides/making-music.md): From a score (or a prompt asking an LLM to write one) to a finished, reverb-polished audio track — all on one canvas, no external software. This page walks through the four stages — **composition → performance → synthesis → mixing** — explaining the musical concepts and every parameter on the way.
- [Model files for the shipped workflows](https://comfytv.org/guides/models.md)
- [Panorama (360°)](https://comfytv.org/guides/panorama.md)
- [Sidebar config editor](https://comfytv.org/guides/sidebar-config-editor.md)
- [Video & audio](https://comfytv.org/guides/video-and-audio.md): Video **generation** lives in the Generate group (see [generate.md](generate.md)). This page covers the **editing** stages in **ComfyTV / Video** and the separation stages in **ComfyTV / Audio**. Video Upscale, Subtitle Erase, and Demucs vocal/bg split are pending — see [roadmap.md](roadmap.md).
- [ComfyTV](https://comfytv.org/index.md): Turn ComfyUI into a TapNow / LibTV-style canvas app. Every operation is a node; results flow downstream automatically.
- [Annotate](https://comfytv.org/node-reference/annotate.md): Draw on-screen boxes and grids, fill or mirror the borders, or scroll the frame — burned into the video.
- [Art FX](https://comfytv.org/node-reference/art-fx.md): Stylizes footage into a cartoon, charcoal, emboss, or halftone look, in one ▶ Run.
- [Load Audio from Asset](https://comfytv.org/node-reference/asset-audio-loader.md): Pick audio from the ComfyTV **project asset library**—reuse Speech/Music Stage output or imports for Video Stage, timelines, or reference audio.
- [Load Image from Asset](https://comfytv.org/node-reference/asset-image-loader.md): Pick an image from the ComfyTV **project asset library** as pipeline input—ideal for reusing generated, edited, or imported results in the current project, not raw `input/` files.
- [Load 3D Model from Asset](https://comfytv.org/node-reference/asset-model-loader.md): Pick a 3D model from your project's asset library and bring it into the pipeline—no upload, no generation.
- [Load Video from Asset](https://comfytv.org/node-reference/asset-video-loader.md): Pick a video from the ComfyTV **project asset library** for clip, demux, upscale, or other downstream steps—reuse in-project media, not raw `input/` files.
- [Music Stage (Audio Stage)](https://comfytv.org/node-reference/audio.md): Generates music or songs from style tags and optional lyrics—instrumental or vocal—not the same as Speech Stage text-to-speech.
- [Audio Analyze](https://comfytv.org/node-reference/audio-analyze.md): Measure a track — integrated loudness (LUFS), peak/mean volume, statistical stats, or silence segments — and get a structured report.
- [Audio Convolve (IR)](https://comfytv.org/node-reference/audio-convolve.md): Convolution reverb — stamp the acoustic character of a real space (or any impulse response) onto your audio.
- [Audio Crossfade](https://comfytv.org/node-reference/audio-crossfade.md): Blend the tail of one clip into the head of the next so cuts don't sound abrupt.
- [Audio Deconvolve (IR)](https://comfytv.org/node-reference/audio-deconvolve.md): Recover an impulse response from a recorded sine-sweep — turn a room, cabinet, or device into a reusable IR.
- [Audio Denoise](https://comfytv.org/node-reference/audio-denoise.md): Strip steady background noise from a track — fan hum, tape hiss, air-con drone — or trim out silent gaps.
- [Audio Duck](https://comfytv.org/node-reference/audio-duck.md): Automatically dip a music or background track whenever a voice (or other sidechain) is present — sidechain compression.
- [Audio Dynamics](https://comfytv.org/node-reference/audio-dynamics.md): Tame or shape the level of a track with a compressor, gate, limiter, or de-esser — the same dynamics tools you'd reach for in a DAW.
- [Audio Echo](https://comfytv.org/node-reference/audio-echo.md): Add echo, slapback, or feedback repeats to a track — from a tight vocal double to canyon-sized reflections.
- [Audio EQ](https://comfytv.org/node-reference/audio-eq.md): A parametric equalizer driven by a graphical EQ curve — boost, cut, and filter frequency bands to shape the tone of a track.
- [Audio · Background Only](https://comfytv.org/node-reference/audio-extract-bg.md): **Roadmap / pending** — Demucs accompaniment / instrumental stem. Pairs with **Vocals Only**; `audio-bg/` reserved.
- [Audio · Vocals Only](https://comfytv.org/node-reference/audio-extract-vocal.md): **Roadmap / pending** — Demucs vocal stem separation from video/mix. Workflow slot exists; backend not wired.
- [Load Audio](https://comfytv.org/node-reference/audio-loader.md): Pick or upload local audio from ComfyUI's `input/` folder as the **starting point** for ComfyTV audio pipelines—voiceover, mixing, and IA2V video inputs.
- [Audio Loudness](https://comfytv.org/node-reference/audio-loudness.md): Bring a track to a consistent, broadcast-safe loudness — EBU R128 normalization, adaptive levelling, or simple peak/RMS/LUFS matching.
- [Audio Meter Overlay](https://comfytv.org/node-reference/audio-meter.md): Burn a live level (VU) meter onto a video that moves with its own audio — great for showing loudness on screen.
- [Audio Beats & Notes](https://comfytv.org/node-reference/audio-mir.md): Analyze a track for its beats, onsets, or note events and turn them into keyframes and labels — so animation can move to the music.
- [Audio Mix](https://comfytv.org/node-reference/audio-mix.md): Layer up to four audio tracks into one, each with its own volume and stereo position — a mini mixing desk.
- [Audio Modulation](https://comfytv.org/node-reference/audio-modulation.md): Classic time-modulation effects — phaser, flanger, chorus, vibrato, tremolo, and pulsator — in one node.
- [Noise Reduction (Spectral)](https://comfytv.org/node-reference/audio-noise-reduction.md): Learn the sound of steady background noise and subtract it, spectrally — like Audacity's Noise Reduction.
- [Audio Picker](https://comfytv.org/node-reference/audio-picker.md): Pick **one** audio track out of a pool of generated clips to send downstream—the audio counterpart of Image Picker.
- [Audio Reactive](https://comfytv.org/node-reference/audio-reactive.md): Turn a frequency band's loudness over time into animation keyframes — so a parameter follows the music's energy.
- [Audio Repair](https://comfytv.org/node-reference/audio-repair.md): Fix specific audio defects — clicks and pops, clipped/distorted peaks, denormals, and mains-frequency hum — with targeted restoration filters.
- [Audio Saturate](https://comfytv.org/node-reference/audio-saturate.md): Add warmth, grit, or lo-fi character — soft-clipping, psychoacoustic clipping, bit-crushing, harmonic excitement, and crystalizer.
- [Audio Split Export](https://comfytv.org/node-reference/audio-segment-export.md): Cut one recording into many separate audio files — either at silences it detects automatically, or at time ranges you specify.
- [Audio Stem Split](https://comfytv.org/node-reference/audio-stem-split.md): Separate a finished mix into its parts — vocals, drums, bass, and everything else — for remixing or clean-up.
- [Audio Stereo](https://comfytv.org/node-reference/audio-stereo.md): Reshape the stereo field — widen, narrow, cross-feed, Haas-delay, balance, collapse to mono, or swap channels.
- [Audio Sweep (ESS)](https://comfytv.org/node-reference/audio-sweep.md): Generate an exponential sine-sweep test signal — the excitation used to capture a room or device's impulse response.
- [Audio Time / Pitch](https://comfytv.org/node-reference/audio-time-pitch.md): Change speed, shift pitch, or time-stretch a track — with a high-quality phase-vocoder option that keeps pitch and tempo independent.
- [Demux · Audio Track](https://comfytv.org/node-reference/audio-video-demux-audio.md): **Extract the audio stream** from a video container as `COMFYTV_AUDIO`. **PyAV** — pairs with **Demux · Silent Video**.
- [Demux · Silent Video](https://comfytv.org/node-reference/audio-video-demux-video.md): Strip the audio track, keep video → silent `COMFYTV_VIDEO`. **PyAV** — pairs with **Demux · Audio Track**.
- [Audio Visualize](https://comfytv.org/node-reference/audio-visualize.md): Render a track as a picture — waveform or spectrogram, in a quick FFmpeg style or a higher-detail "pro" render.
- [← ComfyTV Audio](https://comfytv.org/node-reference/bridge-from-audio.md): **Out-bridge**: **COMFYTV_AUDIO** wav URL → native **AUDIO** dict. **No ▶ Run** — torchaudio load on Queue.
- [← ComfyTV Image](https://comfytv.org/node-reference/bridge-from-image.md): **Out-bridge**: **COMFYTV_IMAGE** URL snapshot → native **IMAGE** tensor. **No ▶ Run** — loads PNG from disk when the ComfyUI graph executes.
- [← ComfyTV Mask](https://comfytv.org/node-reference/bridge-from-mask.md): **Out-bridge**: extract **MASK** tensor from **COMFYTV_IMAGE** (often PNG with alpha). **No ▶ Run** — loads on Queue.
- [← ComfyTV Text](https://comfytv.org/node-reference/bridge-from-text.md): **Out-bridge**: **COMFYTV_TEXT** snapshot → native **STRING**. **No ▶ Run** — passes string on Queue.
- [← ComfyTV Video](https://comfytv.org/node-reference/bridge-from-video.md): **Out-bridge**: **COMFYTV_VIDEO** URL → native **VIDEO** object. **No ▶ Run** — loads mp4 on Queue.
- [→ ComfyTV Audio](https://comfytv.org/node-reference/bridge-to-audio.md): **Into-bridge**: ComfyUI **AUDIO** dict → **COMFYTV_AUDIO** wav URL. **▶ Run** writes `output/ComfyTV/bridge/*.wav`.
- [→ ComfyTV Image](https://comfytv.org/node-reference/bridge-to-image.md): **Into-bridge**: convert native ComfyUI **IMAGE** tensor (any plugin) to a **COMFYTV_IMAGE** URL snapshot. Click **▶ Run** to write `output/ComfyTV/bridge/` and persist in the project.
- [→ ComfyTV Images](https://comfytv.org/node-reference/bridge-to-images.md): **Into-bridge**: save each frame of a ComfyUI **IMAGE batch** as PNG → **COMFYTV_IMAGES** JSON + selected **COMFYTV_IMAGE**. Requires **▶ Run**.
- [→ ComfyTV Text](https://comfytv.org/node-reference/bridge-to-text.md): **Into-bridge**: any ComfyUI **STRING** (prompt enhance, caption, LLM) → **COMFYTV_TEXT** snapshot. **▶ Run** persists in project lineage (text stored as string, not PNG).
- [→ ComfyTV Video](https://comfytv.org/node-reference/bridge-to-video.md): **Into-bridge**: ComfyUI **VIDEO** object (e.g. **mesh2motion**) → **COMFYTV_VIDEO** URL. **▶ Run** writes `output/ComfyTV/bridge/*.mp4`.
- [Card 3D](https://comfytv.org/node-reference/card3-d.md): Treat your footage as a flat "card" floating in 3D space — move it, rotate it, and push it toward or away from the camera with real perspective.
- [CDL Grade](https://comfytv.org/node-reference/cdl.md): An ASC-CDL primary grade — slope, offset, power per channel plus saturation, the portable grade format every finishing suite understands.
- [Chord Accompaniment](https://comfytv.org/node-reference/chord-accomp.md): Type a chord progression and get a played-out backing part — voiced, patterned and exported as MIDI.
- [Chroma Shift](https://comfytv.org/node-reference/chroma-shift.md): Offset the red and blue channels independently for a chromatic-aberration / RGB-split / glitch look.
- [Chromatic Aberration](https://comfytv.org/node-reference/chromatic-aberration.md): Split the red/green/blue channels outward from a center point for that "cheap lens" color fringing look — or dial it negative to pull fringing in.
- [Click Track](https://comfytv.org/node-reference/click-track.md): Generate a metronome click as audio — set a tempo and meter, or build the click from labels.
- [ColorGradeStage](https://comfytv.org/node-reference/color-grade.md)
- [Color Suppress](https://comfytv.org/node-reference/color-suppress.md): Knock down any of six primary/secondary colors independently — a targeted spill and tint remover.
- [CompareStage](https://comfytv.org/node-reference/compare.md)
- [Contact Sheet](https://comfytv.org/node-reference/contact-sheet.md): Lay out evenly-sampled frames from a video into a single grid image — a review "宫格" / contact sheet for scanning a clip at a glance.
- [Corner Pin](https://comfytv.org/node-reference/corner-pin.md): Warp a clip by dragging its four corners — for screen replacements, sign inserts, and perspective fits.
- [CropStage](https://comfytv.org/node-reference/crop.md)
- [CutoutStage](https://comfytv.org/node-reference/cutout.md)
- [Despill](https://comfytv.org/node-reference/despill.md): Strip the green or blue screen color that bleeds onto your subject, without touching the key itself.
- [Director Timeline](https://comfytv.org/node-reference/director-timeline.md): Arrange image clips and optional audio on a visual timeline, outputting structured `COMFYTV_TIMELINE`—the "edit decision list" for multi-shot short video.
- [EraseStage](https://comfytv.org/node-reference/erase.md)
- [Expression](https://comfytv.org/node-reference/expression.md): Turn a math expression into an animation curve — a per-frame value track you can drive an FX parameter with, like a Nuke/AE expression or a wiggle.
- [Face Blur](https://comfytv.org/node-reference/face-blur.md): Automatically detect faces in a clip and obscure them with a blur, pixelation, or box.
- [Feedback FX](https://comfytv.org/node-reference/feedback-fx.md): Recursive video feedback — vertigo zoom, echo trails, or nervous jitter, in one ▶ Run.
- [Frame Blend](https://comfytv.org/node-reference/frame-blend.md): Blend neighboring frames together for motion trails, echo, or motion-blur — like frame-blending / motion blur in a compositor.
- [FX Chain](https://comfytv.org/node-reference/fxchain.md): Bake a whole chain of upstream FX stages into the video in a single transcode pass, and set the delivery format.
- [Glitch FX](https://comfytv.org/node-reference/glitch-fx.md): Datamosh-style digital glitch — block shifts and RGB channel splits, in one ▶ Run.
- [Glow](https://comfytv.org/node-reference/glow.md): Bloom / glow that makes bright areas bleed light — like a lens bloom or After Effects "Glow", with a multi-blur soft falloff.
- [God Rays](https://comfytv.org/node-reference/god-rays.md): Volumetric light shafts (crepuscular rays) that radiate out from a point — like sunbeams through clouds or a window.
- [Auto White Balance](https://comfytv.org/node-reference/gray-world.md): One-click neutral color — removes a color cast automatically using the gray-world assumption.
- [GridSplitStage](https://comfytv.org/node-reference/grid-split.md)
- [Histogram Equalize](https://comfytv.org/node-reference/histogram-eq.md): Redistribute tones to expand contrast in flat footage — with a strength dial and optional CLAHE-style clipping to tame noise.
- [Hue Correct](https://comfytv.org/node-reference/hue-correct.md): Retune specific hues without touching the rest — bend saturation, luminance, or a channel as a function of hue.
- [Image Stage](https://comfytv.org/node-reference/image.md): Generates one or more images from text with local or built-in models—the main “create from scratch” entry in ComfyTV, with img2img (i2i) support.
- [ImageEditStage](https://comfytv.org/node-reference/image-edit.md)
- [ImageLoaderStage](https://comfytv.org/node-reference/image-loader.md)
- [ImagePickerStage](https://comfytv.org/node-reference/image-picker.md)
- [ImageVariationsStage](https://comfytv.org/node-reference/image-variations.md)
- [InpaintStage](https://comfytv.org/node-reference/inpaint.md)
- [Kaleidoscope](https://comfytv.org/node-reference/kaleidoscope.md): Mirrors the frame into radial wedges for a symmetric kaleidoscope pattern, in one ▶ Run.
- [Ken Burns](https://comfytv.org/node-reference/ken-burns.md): Turns a single still image into a video with a slow pan-and-zoom move, in one ▶ Run.
- [Key Mix](https://comfytv.org/node-reference/key-mix.md): Composite one clip over another through a matte — the "put the keyed subject on the new background" step.
- [Keyer](https://comfytv.org/node-reference/keyer.md): A flexible luma/chroma keyer with independent tolerance and softness on the low and high sides — pull mattes from brightness, a color, or a screen.
- [Layer Editor](https://comfytv.org/node-reference/layer-editor.md): A full raster + vector image editor built into the node — layers, masks, selections, brushes, shapes, text and transforms — like a compact GIMP/Photoshop right inside your ComfyTV graph.
- [Lens Distort](https://comfytv.org/node-reference/lens-distort.md): Add or remove barrel/pincushion lens distortion using real camera-lens models — the same math VFX tools like Nuke and 3DEqualizer use.
- [Lens Flare](https://comfytv.org/node-reference/lens-flare.md): Drop a procedural lens flare — glow, streaks, and ghost blobs — anywhere in frame, aimed at a bright light source.
- [Light Graffiti](https://comfytv.org/node-reference/light-graffiti.md): Paints with light — bright pixels leave glowing trails across the clip, in one ▶ Run.
- [Line Art](https://comfytv.org/node-reference/line-art.md): Render a 3D model to clean line art — silhouettes, creases, and open borders, with hidden-line removal — ready as a ControlNet lineart image.
- [Make Proxy](https://comfytv.org/node-reference/make-proxy.md): Build a lightweight preview copy of a video so playback and scrubbing stay smooth.
- [Mask Cleanup (ComfyTV)](https://comfytv.org/node-reference/mask-cleanup.md): Removes stray specks and fills pinholes in a mask so downstream cutouts, keys, and inpaints have clean, solid edges.
- [Mask Propagate](https://comfytv.org/node-reference/mask-propagate.md): Take a mask drawn on one frame and carry it across the whole moving clip.
- [Material](https://comfytv.org/node-reference/material.md): A PBR material-ball editor in a node: spin a live WebGL sphere, dial in physically-based surface properties (metal, glass, plastic, ceramic…), and emit both a reusable material and a snapshot preview image.
- [Matte Monitor](https://comfytv.org/node-reference/matte-monitor.md): A QC view that exaggerates the gray zones of a matte so you can spot leaks and soft edges instantly.
- [Matte Morphology](https://comfytv.org/node-reference/matte-morph.md): Grow, shrink, or clean a matte's edges with erode/dilate/open/close — the standard fix for chattery or leaky keys.
- [Mesh Bake Maps](https://comfytv.org/node-reference/mesh-bake-maps.md): Bake a high-poly mesh's detail onto a low-poly one — a tangent-space normal map and/or an ambient-occlusion map — so a light model still reads as if it had all the geometry.
- [Mesh Boolean](https://comfytv.org/node-reference/mesh-boolean.md): Combine two meshes with CSG — union to merge, difference to cut one out of the other, intersect to keep only the overlap — positioned with an on-card gizmo.
- [Mesh Ops](https://comfytv.org/node-reference/mesh-op.md): One node, many mesh operations — decimate, remesh, weld, fill holes, smooth normals, subdivide, UV-unwrap, or export — chosen from a dropdown, with its own parameters shown on the card.
- [Mesh Primitive](https://comfytv.org/node-reference/mesh-primitive.md): Generate a clean base mesh — cube, sphere, cylinder, cone, plane, or torus — to start a 3D build or feed the rest of the geometry workshop.
- [MIDI Editor](https://comfytv.org/node-reference/midi-editor.md): A seconds-domain MIDI piano roll. Draw and edit notes across channels with per-channel instruments, mute/solo and velocity, then export a Standard MIDI File.
- [MirrorStage](https://comfytv.org/node-reference/mirror.md)
- [Load 3D Model](https://comfytv.org/node-reference/model-loader.md): Load a 3D model file from ComfyUI's `input/3d` folder, orbit it in-body, and optionally bind materials onto individual mesh parts.
- [3D Model Stage](https://comfytv.org/node-reference/model3-d.md): Generate a 3D model (GLB) from a prompt, reference images, and/or existing models—your text/image-to-3D generation stage.
- [Motion Track](https://comfytv.org/node-reference/motion-track.md): Click points on the video to track them frame-by-frame, and optionally solve a transform you can pin things to.
- [MultiangleStage](https://comfytv.org/node-reference/multiangle.md)
- [Muse Reverb (FDN)](https://comfytv.org/node-reference/muse-reverb.md): An algorithmic feedback-delay-network reverb with fine control over decay, tone, room size, and stereo spread.
- [Old Film](https://comfytv.org/node-reference/old-film.md): Full aged-film simulation — dust, scratches, brightness flicker, uneven development, and frame weave, like footage run through a worn projector.
- [OutpaintStage](https://comfytv.org/node-reference/outpaint.md)
- [Paint Strokes](https://comfytv.org/node-reference/paint-stroke.md): Paint clone, blur, or color strokes directly on the video to retouch or reveal, over a time window.
- [Panorama](https://comfytv.org/node-reference/panorama.md): Interactive 360° panorama viewer: upload an HDRI / equirectangular image, or generate a full surround scene with an AI workflow, then look around from the center of the sphere.
- [Panorama · Current View](https://comfytv.org/node-reference/panorama-current-view.md): Aim anywhere inside the 360° panorama viewer and capture the current viewport as a flat image (`COMFYTV_IMAGE`).
- [Panorama · Multi-View](https://comfytv.org/node-reference/panorama-multi-view.md): Capture multiple evenly spaced viewport shots from one 360° panorama in a single Run — outputs a `COMFYTV_IMAGES` batch for storyboards and multi-camera setups.
- [Particles](https://comfytv.org/node-reference/particles.md): A 2D particle generator — sparks, embers, smoke, magic dust — composited over your clip, with a live Canvas2D preview.
- [Pattern](https://comfytv.org/node-reference/pattern.md): A procedural clip generator — gradients, noise, color bars, test wheels, countdowns and more, in one ▶ Run.
- [PIK Keyer](https://comfytv.org/node-reference/pik.md): An IBK-style image-based primatte keyer for hard shots — feed it a clean plate and garbage masks for edges even a simple chroma key can't hold.
- [Posterize](https://comfytv.org/node-reference/posterize.md): Reduce the clip to a small palette of flat colors — a poster / screen-print / cel-shaded look via color quantization.
- [Project](https://comfytv.org/node-reference/project.md): Names and binds the ComfyTV project context so every stage output is stored under one project and restored when you reload the workflow.
- [Pseudocolor](https://comfytv.org/node-reference/pseudocolor.md): Map the picture's brightness onto a color palette — false-color / heatmap looks, or scientific colormaps like viridis and turbo.
- [Regrain](https://comfytv.org/node-reference/regrain.md): Add realistic film grain with separate control over shadows, midtones, and highlights — for regraining a clean render or matching real film stock.
- [RelightStage](https://comfytv.org/node-reference/relight.md)
- [RotateStage](https://comfytv.org/node-reference/rotate.md)
- [Roto Mask](https://comfytv.org/node-reference/roto-mask.md): Draw a spline shape on the video and bake it into an animated mask, right on the node.
- [Scene Detect](https://comfytv.org/node-reference/scene-detect.md): Find the cuts in a video and split it into per-shot thumbnails or ready-to-use clips.
- [3D Scene](https://comfytv.org/node-reference/scene3-d.md): A DCC-style 3D workbench inside a node: build a scene from characters, primitives, imported models, lights and cameras, then capture a still, a multi-camera batch, or record a video — in color, depth, normal, or OpenPose channels.
- [Score](https://comfytv.org/node-reference/score.md): Turn MusicXML into an engraved, playable sheet-music score you can pass down the music lane.
- [Score Editor](https://comfytv.org/node-reference/score-editor.md): A beat-domain piano-roll editor. Draw, step-enter and edit notes across up to four parts, then export a MusicXML score.
- [Score Performer](https://comfytv.org/node-reference/score-to-midi.md): Turn a written score into an expressive performance — swing, human feel, per-instrument phrasing — and get a MIDI file out.
- [Select0r](https://comfytv.org/node-reference/select0r.md): A color-region keyer that carves a matte out of a chosen color space with tunable shape and edge falloff.
- [Selective Color](https://comfytv.org/node-reference/selective-color.md): Photoshop-style Selective Color — adjust one color family at a time without a mask.
- [Sequence](https://comfytv.org/node-reference/sequence.md): Assemble multiple clips end-to-end into one edit, with per-cut in/out trims and optional transitions.
- [Score Synth](https://comfytv.org/node-reference/sf2-synth.md): Render a performance to real audio with a SoundFont — ComfyTV's built-in SF2/SF3 synthesizer.
- [Shape Mask](https://comfytv.org/node-reference/shape-mask.md): Generate a matte from a procedural gradient shape — radial, linear, clock wipe, checker and more — or from a luminance image.
- [Shot Images](https://comfytv.org/node-reference/shot-images.md): Automatically run an image workflow **once per shot** from a **Storyboard**—turn a text shot list into a visual storyboard image set.
- [Slit Scan](https://comfytv.org/node-reference/slit-scan.md): Time-smears a clip so different rows/columns come from different moments — the classic slit-scan look, in one ▶ Run.
- [Speech Stage](https://comfytv.org/node-reference/speech.md): Synthesizes spoken narration and dialogue (TTS) from a script—not the same as Music Stage song/instrumental generation.
- [Split Parts](https://comfytv.org/node-reference/split-part.md): Segment an image into separate part cutouts — click or box the pieces you want (a strap, a buckle, a logo), or name them by text — and get each one back as its own image.
- [Spot Remover](https://comfytv.org/node-reference/spot-remover.md): Cover a fixed rectangular spot — a logo, wire, or blemish — by blending or inpainting it away.
- [UV Remap](https://comfytv.org/node-reference/stmap.md): Warp a clip by an ST-map / distortion map — the standard way to re-apply lens distortion, screen warps, or any per-pixel displacement.
- [STMap Generate](https://comfytv.org/node-reference/stmap-gen.md): Bake a lens-distortion model into an STMap image — a reusable UV displacement map you can apply anywhere, in ComfyTV or any tool that reads STMaps.
- [Storyboard Stage](https://comfytv.org/node-reference/storyboard.md): Uses an LLM to expand a short story premise into a structured shot list (16 fields per shot) for Shot Images—this node does not generate pictures itself.
- [Storyboard Editor](https://comfytv.org/node-reference/storyboard-editor.md): A shot-board workbench — draw and annotate every panel with the full Layer Editor engine, set timing, onion-skin between shots, play the animatic, and export to video/GIF/PDF/ZIP/PSD.
- [Strobe](https://comfytv.org/node-reference/strobe.md): Replaces frames on a regular interval with black, white, or a held frame for a strobing pulse, in one ▶ Run.
- [Subtitles](https://comfytv.org/node-reference/subtitle.md): Burn timed caption cues from SRT or WebVTT text onto a video — open captions rendered directly into the picture.
- [Subtitles · Speech-to-Text](https://comfytv.org/node-reference/subtitle-gen.md): Auto-transcribe the speech in a video or audio clip into subtitle text, ready to feed the Subtitles node.
- [Text Stage](https://comfytv.org/node-reference/text.md): Generates text with a local LLM from your instructions—expand prompts, write descriptions, or feed context to image and video stages.
- [Input Text](https://comfytv.org/node-reference/text-loader.md): Type or paste text and feed it into your pipeline as a `COMFYTV_TEXT`—the simplest way to author prompts, captions, or notes on the canvas.
- [Time Remap](https://comfytv.org/node-reference/time-remap.md): Variable-speed retiming and freeze-frames — ramp a clip from slow-mo to fast, or hold on a single frame.
- [Timeline Render](https://comfytv.org/node-reference/timeline-video.md): Encode a **Director Timeline** sequence into a finished video—the **export** step of a multi-shot timeline workflow.
- [Title](https://comfytv.org/node-reference/title.md): Burn a styled text title card onto a video — like a lower-third or opening title, with fonts, stroke, placement, timing, fade and typewriter reveal.
- [UpscaleStage](https://comfytv.org/node-reference/upscale.md)
- [Video Stage](https://comfytv.org/node-reference/video.md): Generates short video clips from text, images, and optional audio—four built-in LTX 2.3 modes: T2V, I2V, FLF2V, and IA2V.
- [Blur / Sharpen](https://comfytv.org/node-reference/video-blur-sharpen.md): Softens or sharpens footage with a choice of blur kernels or an unsharp mask, in one ▶ Run.
- [Chroma Key](https://comfytv.org/node-reference/video-chroma-key.md): Pull a green- or blue-screen key out of a clip with built-in spill suppression, in one pass.
- [Video Clip](https://comfytv.org/node-reference/video-clip.md): Trim a video to a [start_s, end_s] range, keep original frame rate. **PyAV** on disk → new `COMFYTV_VIDEO` snapshot.
- [Video Color](https://comfytv.org/node-reference/video-color.md): A full primary color-corrector for clips — exposure, white balance, hue/sat, levels and three-way color wheels in one node.
- [Video Composite](https://comfytv.org/node-reference/video-composite.md): Layer one video over another with a blend mode, opacity, a 2D transform, an optional mask, and animatable keyframes.
- [Video Concat](https://comfytv.org/node-reference/video-concat.md): Join two or more clips end-to-end into a single continuous video, in an order you control.
- [Video Crop](https://comfytv.org/node-reference/video-crop.md): Crop the same rectangle on every frame. **PyAV** processing, frame rate unchanged → `COMFYTV_VIDEO`. This is a **PyAV media-processing** stage—not AI video generation via **Video Stage**.
- [Video Curves](https://comfytv.org/node-reference/video-curves.md): Photoshop-style tone curves for a clip — master and per-channel, plus ready-made film-look presets.
- [Deinterlace](https://comfytv.org/node-reference/video-deinterlace.md): Converts interlaced footage to clean progressive frames, removing the comb/tear artifacts.
- [Video Denoise](https://comfytv.org/node-reference/video-denoise.md): Cleans up grain, sensor noise, and banding across a whole clip with a choice of denoise algorithms.
- [Extract Frame](https://comfytv.org/node-reference/video-extract-frame.md): Grab one still from a video (first, last, middle, or a custom timestamp) as `COMFYTV_IMAGE`. Uses **PyAV** on disk — no GPU.
- [Extract Frames](https://comfytv.org/node-reference/video-frames.md): Grab still frames from a clip at timestamps you mark, and get them back as an image batch.
- [Frame Interpolate](https://comfytv.org/node-reference/video-interpolate.md): Generates in-between frames to raise the frame rate or create smooth slow motion.
- [Load Video](https://comfytv.org/node-reference/video-loader.md): Pick or upload a local video from ComfyUI's `input/` folder as the **starting point** for ComfyTV video pipelines—clip, extract frames, demux, and more.
- [Luma Wipe](https://comfytv.org/node-reference/video-luma-wipe.md): Transition between two clips using a grayscale gradient pattern — the classic "gradient wipe" reveal.
- [Video LUT](https://comfytv.org/node-reference/video-lut.md): Apply a finished color look to a clip from a `.cube` 3D LUT in your library.
- [Mux Audio](https://comfytv.org/node-reference/video-mux-audio.md): Attach an audio track to a clip — either replacing the original sound or mixing on top of it.
- [Video Picker](https://comfytv.org/node-reference/video-picker.md): Pick **one** clip out of a pool of generated videos to send downstream—the video counterpart of Image Picker.
- [Video Resize](https://comfytv.org/node-reference/video-resize.md): Change output width/height (-1 keeps aspect). Frame rate unchanged. **PyAV** resample → `COMFYTV_VIDEO`. This is a **PyAV media-processing** stage—not AI video generation via **Video Stage**.
- [Video Rotate](https://comfytv.org/node-reference/video-rotate.md): Rotate a clip in 90° steps and/or mirror it horizontally or vertically.
- [Video Scopes](https://comfytv.org/node-reference/video-scopes.md): Render a broadcast scope — waveform, RGB parade, vectorscope, histogram or CIE — for one frame of your video, to check exposure and color.
- [Video Speed](https://comfytv.org/node-reference/video-speed.md): Speed a clip up or slow it down (with optional reverse), keeping the audio in tune.
- [Video Split](https://comfytv.org/node-reference/video-split.md): Cut one clip into two pieces at a single point in time.
- [Video Stabilize](https://comfytv.org/node-reference/video-stabilize.md): Fast single-pass shake reduction using ffmpeg's deshake filter.
- [Stabilize Pro](https://comfytv.org/node-reference/video-stabilize-v2.md): Two-pass camera-path stabilization with adjustable smoothing and auto-zoom to hide borders.
- [Video Stylize](https://comfytv.org/node-reference/video-stylize.md): A grab-bag of one-click stylization looks — vignette, film grain, pixelize, edge detect, sepia, monochrome, or a combined old-film treatment.
- [Subtitle Erase · Region](https://comfytv.org/node-reference/video-subtitle-select-erase.md): **Roadmap / pending** — erase subtitles inside a user-drawn rectangle per frame. Box UI incomplete; Run raises `StageNotImplemented`.
- [Subtitle Erase · Smart](https://comfytv.org/node-reference/video-subtitle-smart-erase.md): **Roadmap / pending** — auto-detect and remove burned-in subtitles. Run raises `StageNotImplemented`.
- [Video Transform](https://comfytv.org/node-reference/video-transform.md): Pan, scale, rotate and skew a clip in 2D, with optional keyframe animation and motion blur.
- [Video Transition](https://comfytv.org/node-reference/video-transition.md): Join two clips with an ffmpeg xfade transition — fades, dissolves, slides, wipes, and dozens more.
- [Video Upscale](https://comfytv.org/node-reference/video-upscale.md): **Roadmap / pending** — intended AI per-frame video enlargement (2x / 4x). Run currently raises `StageNotImplemented`; no real backend yet.
- [Video Volume](https://comfytv.org/node-reference/video-volume.md): Adjust a clip's audio level and add fade-in / fade-out at the head and tail.
- [360 Projection](https://comfytv.org/node-reference/video360.md): Reproject 360°/equirectangular footage — turn a full panorama into a flat "little planet" view, unwrap a fisheye, or re-aim the virtual camera.
- [360 Stabilize](https://comfytv.org/node-reference/video360-stabilize.md): Smooth out camera shake in 360°/equirectangular footage by damping rotation across the whole sphere.
- [Water](https://comfytv.org/node-reference/water.md): Simulates a rippling fluid surface over the footage — rain drops and swirl currents, in one ▶ Run.
- [Wave Warp](https://comfytv.org/node-reference/wave-warp.md): Ripples the image with a traveling sine wave for a watery/heat-haze distortion, in one ▶ Run.
- [Z Defocus](https://comfytv.org/node-reference/zdefocus.md): Depth-based defocus — blur your footage by distance using a depth map, so foreground or background goes soft with real bokeh discs.