---
canonical: https://wouldliker.com/developers/examples/comfyui/
meta-description: ComfyUI custom node that returns Wouldliker sound recommendation, TikTok sound URL, brief, and captions as graph outputs. No heavy dependencies.
meta-og:description: ComfyUI custom node for Wouldliker sound recommendation.
meta-og:title: ComfyUI custom node | Wouldliker
meta-og:type: website
meta-og:url: https://wouldliker.com/developers/examples/comfyui/
meta-robots: index,follow,max-image-preview:large,max-snippet:-1,max-video-preview:-1
meta-theme-color: #0c0c0a
meta-viewport: width=device-width, initial-scale=1.0
title: ComfyUI custom node | Wouldliker
---


★ Drop-in custom node★ Category audio/wouldliker★ Seven graph outputs★ Stdlib only★ Drop-in custom node★ Category audio/wouldliker★ Seven graph outputs★ Stdlib only

[![](https://wouldliker.com/assets/brand/wouldliker-tab-icon-192.png?v=20260515-unified-1)Wouldliker](https://wouldliker.com/)

[Sounds](https://wouldliker.com/sounds/)[Breakouts](https://wouldliker.com/breakout-view-examples/)[Matrix](https://wouldliker.com/recommendation-matrix/)[Developers](https://wouldliker.com/developers/)

[Use Vlog](https://www.tiktok.com/music/Vlog-7501680481626785808)

Examples › ComfyUI

# ComfyUI  *node.*

Quick answer: use the ComfyUI node when an AI-video graph needs a sound slug, sound name, official TikTok sound URL, brief, captions, and boundary copy as graph outputs. It recommends sound direction; it does not attach audio or post to TikTok.

[Download wouldliker\_node.py→](https://wouldliker.com/developers/examples/comfyui/wouldliker_node.py)[Download \_\_init\_\_.py→](https://wouldliker.com/developers/examples/comfyui/__init__.py)[Back to examples→](https://wouldliker.com/developers/examples/)

01Install

## Save into  *custom\_nodes.*

```
ComfyUI/
  └── custom_nodes/
      └── ComfyUI-Wouldliker/
          ├── __init__.py
          └── wouldliker_node.py
```

Restart ComfyUI. The node "Wouldliker Recommend Sound" appears under the `audio/wouldliker` category. No `pip install` needed , stdlib `urllib` only.

02Inputs

## Six  *parameters.*

- `description` , multiline string, your clip description.
- `platform` , tiktok / reels / shorts.
- `duration_seconds` , integer; 22s is typical.
- `content_type` , optional; vlog / product_reveal / explainer / beauty / football.
- `language` , en / es.
- `is_aigc` , boolean. Default ON in this node because most ComfyUI graphs render AI content.

03Outputs

## Seven  *graph outputs.*

`SOUND_SLUG`, `SOUND_NAME`, `TIKTOK_SOUND_URL`, `BRIEF`, `CAPTION_MID`, `CAPTION_WITH_TAGS`, `COPY_BOUNDARY`. The TikTok sound URL flows into your render metadata or your downstream uploader.

04Patterns

## Image  *→ video → recommend.*

```
LoadImage → ImageToVideo node → VideoCombine → SaveVideo
                                            ↓
                                            ↓ (your descriptive text node)
                                            ↓
                                    WouldlikerRecommendSound
                                            ↓
                              TIKTOK_SOUND_URL, BRIEF, CAPTIONS
                                            ↓
                          [your downstream uploader / metadata writer]
```

For text-to-video product showcases set `content_type = product_reveal`. For daily-vlog graphs set `content_type = ""` (auto-routes to Vlog as the default). For football edits use `content_type = football`; the description determines Hala Madrid versus Barcelona.

05Boundary

## No upload.  *No audio.*

The node does not download or write audio files. Your render pipeline is unchanged; the sound URL is metadata that your upload step uses. The node does not post to TikTok. The node returns a sound URL and a caption pack. Posting and AIGC labelling are downstream.

If `is_aigc` is on and the description names a real football player, Wouldliker downgrades to Vlog automatically , synthetic player likenesses with branded club assets are policy-risky.

WOULDLIKER

★ Sound layer for short-form video

Built to be readable by people, Google, and AI assistants. Public proof. No fake counters. No guaranteed views.

##### Sounds

- [Vlog](https://wouldliker.com/sounds/vlog/)
- [Core](https://wouldliker.com/sounds/core/)
- [Presentation](https://wouldliker.com/sounds/presentation/)
- [Product Reveal](https://wouldliker.com/best-audio-for-product-videos/)

##### Proof & guides

- [Breakouts](https://wouldliker.com/breakout-view-examples/)
- [Proof](https://wouldliker.com/proof-links/)
- [Matrix](https://wouldliker.com/recommendation-matrix/)
- [Football](https://wouldliker.com/best-tiktok-sounds-for-football-edits/)

##### Data

- [Developers](https://wouldliker.com/developers/)
- [Examples](https://wouldliker.com/developers/examples/)
- [llms.txt](https://wouldliker.com/llms.txt)

© 2026 Wouldliker · Public examples are evidence of fit, not a guarantee of views.