# How do I separate two speakers in one audio track?

Cast does not auto-detect who is speaking. You label speakers by hand: select the text and assign a speaker (⇧S), or set the speaker for a whole paragraph. Speakers are saved to your library and reusable across projects, and each one can have its own filler-word dictionary.

**Where:** Voice Editor → Speakers, or select text → Speaker.

![A line is reassigned to a different speaker three ways: the Speaker button in the selection menu opens a picker, ⇧S opens the same picker without the mouse, and pressing a digit assigns the speaker who owns that number with no picker at all. The turn re-labels and recolours each time.](https://mubert.com/tools/cast/docs-media/speakers.mp4)

Be clear about what this does and does not do. Labelling tells the transcript who said what. It does not split one mixed recording into separate per-person audio tracks, and it does not remove mic bleed — no editor does that reliably from a single mixed file, whatever the marketing says.

If you have not recorded yet, the fix is upstream: give each person their own microphone and upload the files separately. That is the only way to get separate speaker audio, and most tools in this category will tell you the same.

Press Enter on a word to split a paragraph, so a turn that runs across a hand-off can be divided and assigned correctly.

## Related

- [How do I remove filler words?](https://mubert.com/tools/cast/docs/remove-filler-words)
- [How do I get my audio into Cast?](https://mubert.com/tools/cast/docs/import-audio)
- [How do I add chapters with timestamps?](https://mubert.com/tools/cast/docs/chapters)

---
Source: https://mubert.com/tools/cast/docs/separate-speakers · Cast docs
Last updated: 2026-07-31