ComfyUI-AV-Handles

ComfyUI-AV-Handles

https://github.com/user-attachments/assets/79817bcc-6f69-490e-99e9-a84ed106c597

Result: 6 Handle frames added → WAN S2V processing → Handles trimmed = Perfect sync, stable output from frame 1 ✓

---

What does it do?

Audio/Video handle management for ComfyUI workflows

Add and remove stabilization frames with synchronized audio for video diffusion models.

Video diffusion models (WAN S2V, WAN Fun Control, LTX2, etc.) often need a few frames to stabilize before producing quality output. This node pack lets you:

• Add repeated first frames as "handles" before your sequence
• Sync audio silence to keep A/V perfectly aligned
• Trim handles after processing to restore original length
• Round frame counts to WAN (4n+1) or LTX2 (8n+1) compatible values
• Process audio-only with precise timing control via manual FPS

Typical workflow:

24 frames → Add 8 handles (32 frames) → Process → Trim 8 handles → 24 frames

Audio-only workflow:

Audio (2.0s) + FPS:30 → Add handles → Process → Trim handles → Audio restored

Installation

Via ComfyUI Manager (Recommended)

Search for "AV Handles" and install directly.

Manual Installation

cd ComfyUI/custom_nodes
git clone https://github.com/pizurny/ComfyUI-AV-Handles.git
# Restart ComfyUI

Alternative: Direct Download

• Download this repository as ZIP
• Extract to ComfyUI/custom_nodes/ComfyUI-AV-Handles/
• Ensure folder structure:

ComfyUI-AV-Handles/
├── nodes/
│   ├── av_handles_add.py
│   └── av_handles_trim.py
├── utils/
│   └── wan_utils.py
└── __init__.py

• Restart ComfyUI

Usage

Nodes will appear under: video/handles

AV Handles Add

Adds handle frames by repeating the first frame + audio silence.

Required Inputs:

• handle_frames (INT) - Frames to add (default: 8, range: 0-100). Set to 0 with a padding_mode enabled for auto mode.

Optional Inputs:

• images (IMAGE) - Input image batch (optional for audio-only)
• audio (AUDIO) - Audio to sync
• padding_mode (ENUM) - Frame count rounding: disabled, WAN (4n+1), LTX2 (8n+1)
• manual_fps (FLOAT) - Manual FPS override (default: 0 = auto-detect, range: 0-120)

Outputs: images, audio, total_frames, handles_added, info

Note: Connect handles_added output to Trim node's handle_frames input for automatic sync (essential when using WAN/LTX2 rounding or auto mode).

AV Handles Trim

Removes handle frames + audio silence from beginning.

Required Inputs:

• handle_frames (INT) - Frames to remove (default: 8, range: 0-100)

Optional Inputs:

• images (IMAGE) - Image batch with handles (optional for audio-only)
• audio (AUDIO) - Audio to trim
• manual_fps (FLOAT) - Manual FPS override (default: 0 = auto-detect, range: 0-120)

Outputs: images, audio, remaining_frames, info

Example Workflows

Basic Video Processing

Load Images (24 frames)
         ↓
AV Handles Add (handle_frames: 8)
         ↓
AnimateDiff or Video Model (32 frames)
         ↓
AV Handles Trim (handle_frames: 8)
         ↓
Save Output (24 frames)

Video + Audio Sync

Load Video (60 frames) + Load Audio (2.0s)
         ↓
AV Handles Add (handle_frames: 12, audio connected)
Output: 72 frames, 2.4s audio
         ↓
Process
         ↓
AV Handles Trim (handle_frames: 12, audio connected)
Output: 60 frames, 2.0s audio ✓

Audio-Only Processing

Load Audio (3.0s)
         ↓
AV Handles Add (handle_frames: 24, manual_fps: 30.0, audio connected, NO images)
Output: Audio with 0.8s silence added (3.8s total)
Info: "Audio-only mode | Handle frames: 24 | FPS: 30.00"
         ↓
Process Audio (effects, filters, etc.)
         ↓
AV Handles Trim (handle_frames: 24, manual_fps: 30.0, audio connected, NO ima