CleanStream

CleanStream filters unwanted speech in live OBS audio for streamers and people recording spoken content. It runs speech recognition on your machine with Whisper through whisper.cpp, and introduces a playback delay while it decides which audio to keep.

• Word filtering: Target filler words such as "uh" and "um", profanity, or other words you choose to block.
• Beep or silence: Replace detected unwanted speech with a beep or silence while keeping the original audio for other segments.
• OBS audio sources: Filter any OBS source that produces audio, so you can use it beyond a microphone feed.
• Adaptive delay: The playback delay adjusts to processing speed to keep filtering decisions aligned with the audio listeners hear.

CleanStream is open source under GPL-2.0 and runs on macOS, Windows and Ubuntu Linux. macOS builds support Intel and Apple Silicon. Windows and Linux builds can use NVIDIA GPUs through CUDA or AMD GPUs through ROCm; both platforms also support Vulkan acceleration.

Whisper inference runs locally. The plugin needs a compatible OBS release, and older OBS releases aren't supported.