Strip filler words automatically

Filler removal runs between transcription and LLM enhancement. It uses a hand-tuned list of verbal filler words and conservative regex rules: it only removes things it is sure about, never rewrites the meaning.

Supported languages

• English (um, uh, basically, like, you know, I mean, sort of, kind of)
• Italian (cioè, tipo, praticamente, ecco, insomma)
• Spanish (eh, este, o sea, pues)
• French (euh, ben, du coup, en fait)
• German (ähm, halt, also, sozusagen)
• Portuguese (tipo, né, ahn, então)

Turning it on

1. 1Open Settings → General.
2. 2Toggle Filler removal on.
3. 3Save.

Works without the LLM

Even with the LLM pass set to Off, filler removal still runs. Useful when you want clean text without paying for a cloud LLM.