The ultimate video editing suite for ComfyUI — edit with natural language or hands-on manual controls.
Use AI to describe edits in plain English, or take full manual control with the Effects Builder and text presets — no LLM required.
Features • Examples • Installation • Quick Start • Prompt Guide • Skills • LLM Setup • Troubleshooting • Contributing • Changelog
See CHANGELOG.md for the full detail.
- 🏔️ Marigold V2 — depth, surface normals and albedo from a frozen Qwen-Image-Edit-2509 DiT plus a per-task LoRA, VAE and baked conditioning. One deterministic Euler step, no CFG, no seed. Reached through
depth (v2)/normals (v2)/albedo (v2)on the existingmarigoldmode, so nothing already in your workflows changes; also selectable as adepth_backendon the Shader Overlay node. ~22.5 GB for the first task, ~2 GB per task after (the base is shared), at roughly 3–4 s per frame - 🎚️ Fixes ComfyUI's stock Marigold V2 sampling — the shipped blueprints patch
ModelSamplingAuraFlowwithsampling="flow", which hands the DiT a half-scale latent and returns0.5·(z − v). This backend defaults toIMG_TO_IMG_VELOCITY, which yieldsz − vfrom an unscaled latent and matches the reference implementation exactly;flowstays selectable to reproduce the stock template - 🔧 BlockSwap block discovery — the swap budget was sized from
model.diffusion_model.blocksonly, so any model naming its stack otherwise fell back to an estimate scaled by Wan's 40-block depth. It now also findstransformer_blocks,double_blocksandsingle_blocks, which matters for 60-block Qwen-Image
📋 Previous Releases
| Version | Highlights |
|---|---|
| v2.20.0 | SCAIL-2 pose-driven character animation, FlashVSR one-step super-resolution, SVI 2.0 Pro infinite-length video, Wan-Animate, MatAnyone2 matting, Meta Sapiens2, SHARP, PhyFPS, multi-source comparison, corrected BT.709 video levels |
| v2.19.0 | DreamID-Omni talking-head generation, FaceCam camera control, Fish Speech TTS, Foundation-1 music samples, Frame Picker node, 15 new GLSL shaders (70 total) |
| v2.18.0 | Kiwi-Edit AI video editing, SAM3 + Kiwi-Edit, RTX Video Super Resolution, SeedVR AI Upscaling, FacePoke expression presets |
| v2.17.0 | FacePoke interactive face editor, driving video reference, shader effects system, Flux Klein FP8, onion skin compositing, unified audio output mode |
| v2.16.0 | ACE-Step AI music generation, SAM-Audio source separation, Video Editor v2 (10 panels), AudioX vocal enhancement, NormalCrafter, Video Depth Anything |
| v2.15.0 | MiniMax-Remover, 5 new no-LLM modes, auto-VRAM tile sizing, VRAM management overhaul |
| v2.14.0 | Video Editor NLE node with timeline, razor, crop, transitions, text overlays, keyboard shortcuts |
| v2.13.0 | AI Background Removal (BRIA RMBG), FLUX Klein toggle, Edit FFmpeg fallback, smarter defaults |
| v2.12.0 | AI Face Animation (LivePortrait), MMAudio in-process inference, MCP progressive disclosure, LaMa safetensors conversion |
| v2.11.0 | MMAudio in-process migration, generate_audio no-LLM mode, MCP tools, CLI binary caching |
| v2.10.0 | FLUX Klein in-process migration, interactive mask drawing UI, model output caching |
| v2.9.1 | AI object removal & editing (FLUX Klein 4B), AI audio generation (MMAudio), AI lip sync (MuseTalk), modular architecture refactor |
| v2.8.0 | Effects Builder node, manual no-LLM mode, text node presets, SAM3 subprocess isolation, 15+ bug fixes |
| v2.7.0 | SAM3 auto-mask & greenscreen, LaMa inpainting, programmatic tool calling (PTC) |
| v2.6.5 | Whisper auto-transcription, karaoke subtitles, whisper model/device controls |
| v2.6.0 | HandlerResult contract, compose decomposition, TextInput node, PiP audio mixing, CLI retry |
| v2.5.0 | PiP borders, Ollama VL auto-embedding, overlay animation delegation |
| v2.4.0 | Zero-memory image paths, pipeline chaining fixes, handler module extraction |
| v2.3.0 | Token usage tracking, LUT color grading, vision system, audio analysis |
| v2.2.0 | 200 skills, dynamic input slots, 48 new skills across all categories |
| v2.0.0 | Dynamic input slots, concat, xfade, split screen, animated overlays, text overlays |
| v1.9.0 | Claude CLI, Cursor Agent CLI, Qwen CLI connectors, 50+ context menu presets |
📄 See CHANGELOG.md for the complete version history.
NotebookLM Overview: Exploring the features and capabilities of ComfyUI-FFMPEGA. (Click to watch on YouTube)
|
Describe edits in plain text: "Make it cinematic with a fade in", "Speed up 2x", "VHS look with grain". The AI agent interprets your prompt and builds the FFMPEG pipeline automatically. |
Use the Effects Builder to visually compose up to 5 effects with parameters. Add text overlays and subtitles via preset-powered Text nodes. Full editing control with zero LLM dependency. |
|
Works with Ollama (local, free) and CLI tools (Gemini CLI, Claude Code, Cursor Agent, Qwen Code). Use any local model — Llama 3.1, Qwen3, Mistral, and more. Or skip the LLM entirely. |
200+ video editing skills across visual effects, audio processing, spatial transforms, temporal edits, encoding, cinematic presets, vintage looks, social media, creative effects, text animations, editing & composition, audio visualization, multi-input operations, transitions, concat, split screen, and AI-powered skills (DreamID-Omni talking-head generation, FaceCam camera control, Fish Speech TTS, Foundation-1 music samples, ACE-Step music generation, SAM-Audio source separation, AudioX vocal enhancement, Whisper transcription, SAM3 masking, MiniMax-Remover object removal, MMAudio generation, MuseTalk lip sync, LivePortrait face animation, NormalCrafter surface normals, Video Depth estimation, AI Upscaling, Marigold dense vision). |
|
26 built-in Effects Builder presets and 10 Text node presets with example content. Save/load/delete your own custom presets. One-click clear to reset. |
Process multiple videos with the same instruction. Generate quick low-res previews before committing to full renders. Quality presets from draft to lossless. |
Worked examples: prompts, the pipelines they produce, and the resulting output.
- ComfyUI (latest)
- Python 3.10+
- FFMPEG installed and in PATH (install guide)
- Node.js 18+ (required for CLI tools: Gemini CLI, Claude CLI, Qwen CLI — download)
- Ollama (optional, for local LLM inference — download)
- Open ComfyUI Manager
- Search for
ComfyUI-FFMPEGA - Click Install
Linux / macOS
cd /path/to/ComfyUI/custom_nodes
git clone https://github.com/AEmotionStudio/ComfyUI-FFMPEGA.git
cd ComfyUI-FFMPEGA
pip install -r requirements.txtWindows (PowerShell)
cd C:\path\to\ComfyUI\custom_nodes
git clone https://github.com/AEmotionStudio/ComfyUI-FFMPEGA.git
cd ComfyUI-FFMPEGA
pip install -r requirements.txtNote: Use whichever Python package manager your ComfyUI venv uses (
pip,uv pip, etc.). The above commands assumepipis available in your ComfyUI virtual environment.
Restart ComfyUI after installation.
- Add an FFMPEG Agent node to your workflow
- Connect a video path or use the input field
- Enter a natural language prompt
- Select your LLM model
- Run the workflow
💡 Tip: Right-click the FFMPEG Agent node to open the FFMPEGA Presets context menu — 200+ categorized effects you can apply with a single click, no prompt typing needed. Great for quick edits or discovering what's available.
| Prompt | What It Does |
|---|---|
"Make it cinematic with a vignette" |
Adds letterbox, color grade, and edge darkening |
"Speed up 2x, keep the audio pitch" |
Doubles speed with pitch-corrected audio |
"Make it look like old VHS footage" |
Adds noise, color shift, scan lines |
"Trim first 5 seconds, resize to 720p" |
Cuts intro and scales down |
"Underwater look with echo on audio" |
Blue tint, blur, and audio echo |
"Pixelate it like an 8-bit game" |
Mosaic/pixel art effect |
"Add 'Subscribe!' text at the bottom" |
Text overlay with positioning |
"Cyberpunk style with neon glow" |
High-contrast neon aesthetic |
"Normalize audio, compress for web" |
Loudness normalization + web optimization |
"Spin the video clockwise" |
Continuous animated rotation |
"Add camera shake" |
Random shake/earthquake effect |
"Fade in from black and out at the end" |
Smooth intro/outro transitions |
"Wipe reveal from the left" |
Directional wipe reveal animation |
"Make it pulse like a heartbeat" |
Rhythmic zoom breathing effect |
"Show audio waveform at the bottom" |
Audio visualization overlay |
"Arrange these images in a grid" |
Multi-image grid collage |
"Create a slideshow with fades" |
Image slideshow with transitions |
"Overlay the logo in the corner" |
Picture-in-picture / watermark |
"Create a side-by-side comparison" |
Video next to image in 2-column grid |
"Create a slideshow starting with the video" |
Video first, then image slides |
"Overlay images in the corners" |
Multiple images auto-placed in corners |
"Split screen, use audio from audio_b" |
Side-by-side video with specific audio track |
"Split screen, mix both audio tracks" |
Side-by-side with both audio tracks blended |
When using an LLM, the AI agent interprets your natural language and maps it to skills with specific parameters. Here's how to get the best results. (For manual editing without an LLM, see the Effects Builder and Text Input sections.)
You can request specific parameter values and the agent will use them directly:
| Prompt | What the Agent Does |
|---|---|
"Set brightness to 0.3" |
brightness:value=0.3 |
"Blur with strength 20" |
blur:radius=20 |
"Speed up to 3x" |
speed:factor=3.0 |
"Crop to 1280x720" |
crop:width=1280,height=720 |
"Deband with threshold 0.3 and range 32" |
deband:threshold=0.3,range=32 |
"CRF 18, slow preset" |
quality:crf=18,preset=slow |
"Fade in for 3 seconds" |
fade:type=in,duration=3 |
- Explicit numbers: "brightness 0.2", "speed 1.5x", "CRF 20" — the agent maps these directly
- Named presets: "VHS look", "cinematic style", "noir" — triggers multi-step preset pipelines
- Chaining operations: "Trim first 5 seconds, resize to 720p, add vignette" — executes in order
- Descriptive goals: "Make it look warmer", "Remove the green screen" — the agent picks the right skills
- Technical terms: "denoise", "deband", "normalize audio" — maps to exact FFmpeg filters
- Vague intensity words: "Make it very blurry" or "a little brighter" — the agent has to guess what number "very" or "a little" means. Tip: use a specific value instead: "blur with radius 15"
- Out-of-range values: Parameters are auto-clamped to their valid range. If you ask for "brightness 5.0" it caps at the max (1.0)
- Complex compositing: Multi-layer effects with precise timing may need to be broken into separate passes
- Format-dependent features: Some effects (like transparency) require specific output formats. H.264/MP4 doesn't support alpha channels
Some skills use AI models that auto-download on first use. You can disable automatic downloads with the allow_model_downloads toggle on the FFMPEG Agent node — runs requiring a missing model will fail with a clear message and a manual download link.
All models are mirrored to first-party AEmotionStudio HuggingFace repos for supply chain resilience. Downloads try the AEmotionStudio mirror first, then fall back to upstream sources.
| Model | Size | Stored In | Triggered By | Manual Download |
|---|---|---|---|---|
| SAM3 (Segment Anything 3) | ~300 MB | ComfyUI/models/SAM3/ |
auto_mask skill, sam3_masking no-LLM mode, Effects Builder SAM3 target |
AEmotionStudio/sam3 — download sam3.safetensors |
| SAM3.1 (Multiplex Tracker) | ~3.5 GB | ComfyUI/models/SAM3.1/ |
Same triggers with sam_version = sam3.1 (default). Video masking runs in-process on ComfyUI's native SAM3 model (VRAM-managed, ~3 GB peak); set FFMPEGA_SAM3_NATIVE=0 to force the legacy subprocess path |
AEmotionStudio/sam3.1 — download sam3.1_multiplex.safetensors |
| Whisper large-v3 | ~3 GB | ComfyUI/models/whisper/ |
auto_transcribe, karaoke_subtitles skills, transcribe / karaoke_subtitles no-LLM modes |
AEmotionStudio/whisper-models |
| Whisper medium | ~1.5 GB | ComfyUI/models/whisper/ |
Same as above (set whisper_model to medium) |
Same as above |
| Whisper small | ~500 MB | ComfyUI/models/whisper/ |
Same as above (set whisper_model to small) |
Same as above |
| Whisper base | ~150 MB | ComfyUI/models/whisper/ |
Same as above (set whisper_model to base) |
Same as above |
| Whisper tiny | ~75 MB | ComfyUI/models/whisper/ |
Same as above (set whisper_model to tiny) |
Same as above |
| LaMa (Large Mask Inpainting) | ~195 MB | ~/.cache/torch/hub/checkpoints/ |
auto_mask:effect=remove (legacy fallback) |
AEmotionStudio/lama-inpainting — download big-lama.safetensors |
| FLUX Klein 4B (Editing/Removal) | ~15 GB (bf16) | ComfyUI/models/flux_klein/ |
auto_mask:effect=remove, auto_mask:effect=edit |
AEmotionStudio/flux-klein |
| MiniMax-Remover (Object Removal) | ~2.5 GB | ComfyUI/models/minimax_remover/ |
auto_mask:effect=remove (when use_minimax_remover is On) |
AEmotionStudio/minimax-remover |
| MMAudio (Video-to-Audio) | ~5.5 GB | ComfyUI/models/mmaudio/ |
generate_audio skill |
AEmotionStudio/mmaudio-models |
| MuseTalk (Lip Sync) | ~1.6 GB (fp16) | ComfyUI/models/musetalk/ |
lip_sync skill |
AEmotionStudio/musetalk-models |
| LivePortrait (Face Animation) | ~497 MB | ComfyUI/models/liveportrait/ |
animate_portrait skill, animate_portrait no-LLM mode |
AEmotionStudio/liveportrait-models |
| Video Depth Anything (Temporal Depth) | ~102–670 MB | ComfyUI/models/video_depth/ |
video_depth no-LLM mode |
AEmotionStudio/video-depth-anything |
| Marigold v1.1 (Dense Vision) | ~2.5 GB per mode | Auto-downloaded by diffusers | marigold no-LLM mode (depth/normals/appearance/lighting) |
AEmotionStudio/marigold-depth-v1-1 |
| Marigold V2 (Dense Vision) | ~22.5 GB first task, ~2 GB each after (20.5 GB base is shared) | ComfyUI/models/{diffusion_models,loras,vae,embeddings}/ |
marigold no-LLM mode, depth (v2) / normals (v2) / albedo (v2); also a depth_backend option on the Shader Overlay node |
Comfy-Org/marigold-v2-0 |
| AI Upscaler (Real-ESRGAN / HAT / DAT / SwinIR) | ~17–170 MB per model | ComfyUI/models/upscale_models/ |
ai_upscale skill, ai_upscale no-LLM mode |
AEmotionStudio/ai-upscale-models |
| BRIA RMBG (rembg) | ~270 MB | ~/.u2net/ |
remove_background skill |
Install with pip install 'comfyui-ffmpega[masking]' — model auto-fetched by rembg |
| ACE-Step 1.5 (Music Generation) | ~5 GB | ComfyUI/models/acestep/ |
ace_step no-LLM mode, generate_music skill |
AEmotionStudio/ACE-Step |
| SAM-Audio (Source Separation) | ~1.2 GB (large) / ~600 MB (fp8) | ComfyUI/models/sam_audio/ |
audio_separate no-LLM mode |
AEmotionStudio/sam-audio |
| AudioX (Vocal Enhancement) | ~1 GB | ComfyUI/models/audiox/ |
AudioX chaining with ACE-Step | AEmotionStudio/audiox |
| NormalCrafter (Surface Normals) | ~2 GB | ComfyUI/models/normalcrafter/ |
normalcrafter no-LLM mode |
AEmotionStudio/NormalCrafter |
| Kiwi-Edit (AI Video Editing) | ~5 GB (FP8) / ~10 GB (BF16) | ComfyUI/models/kiwi_edit_* |
kiwi_edit no-LLM mode |
AEmotionStudio/Kiwi-Edit-Instruct |
| DreamID-Omni (Talking Head) |
~12 GB (FP8) / ~23 GB (BF16) | ComfyUI/models/dreamid_omni/ |
dreamid_omni no-LLM mode |
AEmotionStudio/dreamid-omni |
| FaceCam (Camera Control) | ~16.8 GB (high+low bf16) | ComfyUI/models/diffusion_models/ |
FaceCam node | AEmotionStudio/facecam-wan2.2-14b-bf16 |
| Fish Speech S2 Pro (TTS) | ~6.5 GB (FP8) / ~10.4 GB (BF16) | ComfyUI/models/fish_speech/ |
fish_speech no-LLM mode |
AEmotionStudio/fish-speech-s2-pro |
| Foundation-1 (Music Samples) | ~2 GB | ComfyUI/models/foundation1/ |
foundation1 no-LLM mode |
AEmotionStudio/foundation1-models |
| SeedVR2 (Diffusion Upscaler) | ~3.4 GB (3B INT8) / ~8.3 GB (7B INT8) + ~300 MB (VAE) | ComfyUI/models/SEEDVR2/ or ComfyUI/models/diffusion_models/ (both searched) |
ai_upscale with upscale_model = seedvr2_*. Prefer seedvr2_3b_int8 / seedvr2_7b_int8; raise blockswap_blocks to fit the 7B under 16 GB |
AEmotionStudio/SeedVR2-models (FP8/GGUF). INT8 ConvRot checkpoints are local-only — place seedvr2_{3b,7b}_int8_convrot.safetensors in either folder |
| FlashVSR v1.1 (One-Step VSR) | ~5 GB (DiT + VAE + LQ_proj + TCDecoder) | ComfyUI/models/FlashVSR/ |
ai_upscale with upscale_model = flashvsr_full / flashvsr_tiny / flashvsr_tiny_long |
AEmotionStudio/flashvsr-models — GPL-3.0 |
| SCAIL-2 (Pose-Driven Animation) | ~16 GB (fp8) | ComfyUI/models/diffusion_models/ |
scail2 no-LLM mode. Also needs the Wan 2.1 VAE, UMT5-XXL text encoder and clip_vision_h |
Comfy-Org/SCAIL-2 — wan2.1_14B_SCAIL_2_fp8_scaled.safetensors |
| SVI 2.0 Pro (Infinite-Length Video) | ~500 MB (high + low LoRAs) | ComfyUI/models/svi/version-2.0/ |
svi no-LLM mode. |
AEmotionStudio/svi-loras |
| Wan-Animate (Character Animation) | ~18 GB (DiT) + ~2.5 GB (pose) + ~62 MB (detector) | ComfyUI/models/wan_animate/ |
wan_animate no-LLM mode |
Kijai/WanVideo_comfy_fp8_scaled, Kijai/vitpose_comfy |
| MatAnyone2 (Video Matting) | ~135 MB | ComfyUI/models/matanyone2/ |
video_matting no-LLM mode |
AEmotionStudio/matanyone2 — |
| Sapiens2 (Human-Centric Vision) | ~1 GB (0.4B) – ~10 GB (5B fp16) | ComfyUI/models/sapiens2/<task>/ |
sapiens2 no-LLM mode — pose (308 keypoints), segmentation (29 classes), normals, pointmap, matting |
AEmotionStudio/sapiens2-normal and sibling sapiens2-* repos |
| SHARP (3D Gaussian View Synthesis) | ~400 MB | ComfyUI/models/sharp/ |
sharp no-LLM mode |
AEmotionStudio/sharp — |
| Visual Chronometer (PhyFPS) | ~1 GB | ComfyUI/models/visual_chronometer/ |
phyfps no-LLM mode |
AEmotionStudio/Visual_Chronometer |
Note
Models are only downloaded when you use the corresponding skill for the first time. Core FFmpeg editing skills (200+ of them) require zero model downloads.
Every node, input and output — the Agent node's widgets, the save/load nodes, the Video Editor and the rest.
All 200+ skills by category, their parameters, and how to write your own.
Ollama, the four CLI connectors (Gemini CLI, Claude Code, Cursor Agent, Qwen Code), model choice and troubleshooting.
FFMPEG Not Found
Ensure FFMPEG is installed and in your system PATH:
ffmpeg -versionInstall FFMPEG:
| Platform | Command / Method |
|---|---|
| Ubuntu / Debian | sudo apt install ffmpeg |
| Arch / CachyOS | sudo pacman -S ffmpeg |
| Fedora | sudo dnf install ffmpeg |
| macOS | brew install ffmpeg |
| Windows (winget) | winget install Gyan.FFmpeg |
| Windows (choco) | choco install ffmpeg |
| Windows (scoop) | scoop install ffmpeg |
| Windows (manual) | Download from ffmpeg.org/download, extract, and add the bin/ folder to your system PATH |
Windows PATH tip: After installing, open a new terminal and run
ffmpeg -versionto verify. If not found, you may need to add ffmpeg'sbin/directory to your system PATH manually: Settings → System → About → Advanced system settings → Environment Variables → EditPath.
Ollama Connection Failed
Make sure Ollama is running:
ollama serveIf using a custom URL, set it in the node's ollama_url field.
Model Not Found
Pull the required model first:
ollama pull qwen2.5:8bLLM Returns Empty Response
This usually means:
- The model is still loading (first request after start)
- The prompt is too long for the model's context window
- Try running the same prompt again
- Try a different model
Parameter Validation Errors
FFMPEGA auto-coerces types (float→int) and clamps out-of-range values. If you still see errors, try simplifying your prompt or using a more capable model.
Cancelling a Running Request
If the LLM is taking too long or you want to abort mid-request, close the ComfyUI terminal or restart ComfyUI instead of using the interrupt button. The interrupt button waits for the current LLM response to complete, which can take a while — closing/restarting ComfyUI kills it immediately.
Contributions are welcome! Whether it's bug reports, new skills, or improvements — your help is appreciated.
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/AmazingFeature) - Commit your Changes (
git commit -m 'Add some AmazingFeature') - Push to the Branch (
git push origin feature/AmazingFeature) - Open a Pull Request
This project is licensed under the GPL-3.0 License — see the LICENSE file for details.
Developed by Æmotion Studio
