I wanted a way to watch movies and TV shows from my Plex library with my family without permanently altering the original audio or having to find specially edited versions of everything.
So I built Censorarr.
Censorarr is a free, open-source, self-hosted application that automatically detects configured profanity in movies and TV shows and creates a separate CLEAN audio track with those words muted.
The original audio is left intact, so in Plex you can simply choose which version you want to use.
For example:
English
English - CLEAN
Switch to CLEAN when you want filtered audio, or switch right back to the original whenever you don’t.
What it does
- Uses Whisper speech recognition to locate profanity
- Mutes only the detected portions of the audio
- Adds a separate CLEAN audio stream
- Preserves the original video, audio and subtitle streams
- Automatically processes movies and TV shows
- Can use subtitles as additional evidence to improve detection
- Reprocessing replaces the existing CLEAN track instead of adding duplicates
Plex integration
Censorarr can connect directly to Plex and use it for things such as:
- Media/rating information
- Rating-based filtering
- Avoiding processing media that is currently being played
- Refreshing the Plex library after a CLEAN track is created
Plex is optional, though. Censorarr can also operate directly against normal movie and TV folders.
Other integrations
Optional integrations are available for:
- Radarr
- Sonarr
- Bazarr
It can also run completely standalone.
CPU or GPU
A GPU is not required.
Whisper can run locally on the machine hosting Censorarr using the CPU.
If you have another machine with an NVIDIA GPU, Censorarr also supports a separate GPU Worker, so your NAS/Plex server can handle the media while transcription is performed on another computer.
The GPU Worker can also be installed on the same machine as the main Censorarr installation if your GPU is on the same machine.
Installation
Censorarr currently supports:
- Docker / Docker Compose
- Synology Container Manager
- Native Windows 11 installer
- Native Linux Debian installer
There is a setup wizard in the web interface, and I strongly recommend using Dry Run on a movie or two before letting it modify anything.
GitHub
The project is still very new and under active development.
I built it initially for my own Plex setup, so I’m especially interested in getting it onto some different systems now. If anyone is willing to try it, I’d really appreciate:
- bug reports
- installation/compatibility reports
- suggestions
- feedback on profanity detection
- feature ideas
- pull requests
One important note: speech recognition isn’t perfect, so Censorarr can’t guarantee that every word in every movie will be caught. I’m continuing to work on detection and subtitle-assisted rescue passes to improve that. But, I think it is VERY good so far. I highly recommend having subtitles in the folder as this makes detection quite a bit better. It does’t even have to be subtitles that are in perfect sync, as Censorarr just uses it to find words that Whisper may have missed.
If anyone here has wanted something like this for their Plex library, I’d love to hear what you think.
