Short answer: You can’t swap two different speakers in one pass with Voice Changer. Voice Changer converts one voice into another and doesn’t include speaker separation or per‑speaker voice assignment. (help.elevenlabs.io)
What works today
Use Dubbing Studio to handle multi‑speaker clips:
Create a Dubbing project and check “Create Dubbing Studio.” The dubbing pipeline separates speakers. (elevenlabs.io)
In the Studio timeline, each speaker is on its own track. You can reassign clips between speakers and edit transcripts if needed. (elevenlabs.io)
For each speaker track, click the settings cog and choose the voice you want (Other voices, Track clone, or Clip clone). Assign Speaker 1 to Voice B and Speaker 2 to Voice A to “swap” them. (elevenlabs.io)
Export. Dubbing supports MP4/AAC/SRT/AAF and can export per‑speaker WAV tracks in a ZIP if you need separate stems. (help.elevenlabs.io)
Key limits and costs (so you can pick the right tool)
Voice Changer: max 5 minutes per file; 1,000 credits/minute. (help.elevenlabs.io)
Dubbing: UI up to 45 minutes/500 MB (API up to 2.5 hours/1 GB). Pricing varies by mode; e.g., automatic with watermark 2,000 credits/min, no watermark 3,000, Dubbing Studio 5,000–10,000 credits/min (per 2024‑08‑19 article). (elevenlabs.io)
If you must stay in Voice Changer only: there’s no documented way to assign different target voices to different speakers within one clip; you would need to process each speaker separately in separate runs. (help.elevenlabs.io)
stone obsidian
green kite