If you've followed my previous guide for version 3.2.0, you’ll notice that the NSFW detection method has changed a bit in FaceFusion 3.3.0.
But don’t worry — here's an updated and super simple way to disable the NSFW filter.
You don’t need any coding experience, just follow the steps below carefully.
✅ What This Guide Does
This tutorial helps you disable FaceFusion’s NSFW filter so you can use the tool without content being blurred.
Use responsibly.
🗂️ Step 1: Open the content_analyser.py
File
Go into the
facefusion
folder.Open the file located at:
facefusion/facefusion/content_analyser.py
- If you're using Pinokio, you’ll find it here:
pinokio/api/facefusion-pinokio.git/facefusion/facefusion/content_analyser.py
- Open the file using Notepad, VS Code, or any text editor.
💡 Tip: In Pinokio, you can click "View Folder" in the Files section to navigate directly.
🔍 Step 2: Find the detect_nsfw
Function
Scroll through the file and look for the function that starts like this:
def detect_nsfw
This function controls NSFW detection.
🧠 Step 3: Replace the Code
🔴 Original Code:
def detect_nsfw(vision_frame : VisionFrame) -> bool: is_nsfw_1 = detect_with_nsfw_1(vision_frame) is_nsfw_2 = detect_with_nsfw_2(vision_frame) is_nsfw_3 = detect_with_nsfw_3(vision_frame) return is_nsfw_1 and is_nsfw_2 or is_nsfw_1 and is_nsfw_3 or is_nsfw_2 and is_nsfw_3
✅ Replace It With:
def detect_nsfw(vision_frame : VisionFrame) -> bool: return False
⚠️ Important: Use capital
F
inFalse
.
Python is case-sensitive — usingfalse
will cause an error.
⚠️ Final Tips
Do NOT change the indentation.
Python depends on spacing — if you delete or misalign spaces, the app may crash with anIndentationError
.Use responsibly.
Disabling NSFW filters means you’re bypassing safety checks. Be ethical and mindful of how you use it.
❤️ Support Me
If this guide helped you, please consider supporting me.
Even a small tip helps me keep creating helpful tutorials like this.
Your support means a lot and keeps me motivated to make more beginner-friendly guides. 🙌
Top comments (8)
Amazing. Thanks.
Very cool, super easy to follow steps and I appreciate the heads-up about Python's quirks
Glad it helps.
Bug detected (v3.3.0_NSFW hack):
1) If you load NSFW target image as JPG or JPEG image - Face Enchancer DID NOT apply to final image. (But he show work in console log, and in "Preview" section).
2) If you load target image as WEBP or PNG format - Face Enchancer works.
In default (not hacked) version of FaceFusion - JPG or JPEG target image works well.
***Note: ChatGPT suggest to convert before processing every target image to PNG format (
read_image()
from vision.py) -- via FFMPEG or PIL/Pillow.Or find difference between JPG and PNG in code.
try converting to different file formats and sizes.
mine, even in SFW, some file formats and sizes are problematic.
so, it's an issue from the facefusion itself.
my workaround is to resize or reformat the image.
Or, you can post an issue to official facefusion.
github.com/facefusion/facefusion
Looks like here is official FaceFusion repo on GitHub - but they do not have 'Issue' section, wtf...
Really appreciate how clear you made this for non-coders, and that reminder about indentation in Python is so important. Do you know if this works for future updates too, or does it change each version?
Thank you!
for future updates, I can't guarantee (maybe works, maybe not). But, as long as it open source, hopefully I can find a way to bypass it.