Know where you stand.
Clear filtering status, with recent catches kept secondary. No wall of stats to decode mid-stream.
A little less worry. A lot more you.
For the things you can’t unsay.
Censr helps filter unwanted words before your audio reaches your stream.
Coming to Windows and Linux · Mac after
Filtering is active
Stay in the moment.
Your word list is on the job.
Recent activity3 catches
Just a peek at the controls.
Illustrative controls. This website doesn’t listen to or filter audio.
A quiet companion beside OBS. A quick glance to know what’s happening. Then back to being you.
Illustrative states · always clearly labelled
Clear filtering status, with recent catches kept secondary. No wall of stats to decode mid-stream.
Choose your vocabulary, use a beep or silence, and mute the output when you need a moment.
No forced overlay. No watermark. A small window when you want it, a tray icon when you don’t.
A bit of breathing room between what’s said and what goes out. Censr uses that head start to check your word list and replace a match.
Text illustration only. Actual results depend on recognition, processing time and your audio setup.
“Did you see that
shot?”
Connect it once. Test your route.
Find your flow.
Choose the input you want to filter and a processed output. Route game chat or a call into Censr to include it.
Start with a vocabulary preset, add your own words and phrases, then choose a beep or silence.
Send the processed output to OBS. Make a short local test recording and check that no uncensored route runs alongside it.
Censr’s servers recognize your speech a few seconds at a time, then keep none of it: no recordings, no transcripts. Your word list and your history stay on your computer.
Censr is almost ready.
Windows and Linux first, Mac after.
Unpack it, then run Censr/install.sh to add Censr to your app menu. For now you’ll need an invite code to make an account.
Getting comfortable starts here.
Yes. Censr holds your audio back for a moment, so speech recognition has time to check it. You can set the delay from 1.5 to 5 seconds. Your connection and audio route affect the result; check timing and video sync in a test recording before you go live.
It filters the audio routed into its selected input. To include friends, a call or game chat, you need to route that audio through Censr too. It doesn’t automatically capture every application. A virtual audio cable or your existing mixer may be needed; the installer doesn’t install an audio driver.
No filter can promise that. Recognition mistakes, timing and incorrect routing can let unwanted speech through. This is an early preview, and real audio/OBS acceptance testing is unfinished. Test privately before using it on stream. Censr cannot guarantee compliance with a streaming platform’s rules.
This desktop preview is free to explore. Paid plans, prices and billing are not live. You won’t be asked for payment by this website or this build.
Censr launches on Windows and Linux. macOS comes later.
Both, yes. Censr recognizes speech on its own servers, which send back the words they hear and keep none of them, so it needs an internet connection. It also asks you to sign in. For now, your account is saved on your computer only.