VocalLab Podcast Studio makes multi-speaker episodes: debates, interviews, two-host shows. You either describe what the episode should be about and we write it, or you paste a script you already have.
Either way you end up in the same editor, casting voices and adjusting lines.
Two ways to start
Open Podcast in the Studio sidebar, under Create, and click New episode.
Write it for me — describe the topic, pick a format, name the speakers, choose a length. We plan the episode, write it section by section, and hand you a finished script with the speakers already attributed. This costs points, and the price is shown before you commit.
I have a script — paste text you have already written. We detect the speakers automatically. This is free: no AI is involved until you generate audio.
Writing a script yourself
Put one line per turn, and start each line with the speaker's name and a colon:
ALEX: Welcome back to the show.
JAMIE: Glad to be here — this one's going to be fun.
ALEX: So. Remote work. Did it actually make anyone faster?
A few things worth knowing:
- Names become your cast. Two speakers called the same thing will be merged, so keep them distinct.
- A line without a
NAME:prefix is given to whoever spoke last — handy for long turns split across paragraphs. - Anything before the first speaker label is dropped, so a title block at the top is fine.
- Up to 200,000 characters, which is roughly a four-hour episode.
Casting the voices
When the script is ready you land in the editor with your speakers listed but no voices assigned yet.
Assign a voice to each speaker once, and every one of their lines uses it. Changing a speaker's voice re-voices all their lines together — you never have to go line by line.
Use Auto-cast to fill every empty seat at once, then change the ones you disagree with.
Editing before you generate
The editor is the same one audiobooks use, so everything there applies: reorder lines, split a long turn, add a pause between speakers, set an emotion or a per-line delivery note.
For a conversation, two things are worth the effort:
- Tighten the gaps. Podcast turns run closer together than audiobook narration. Short pauses between speakers make the exchange feel live.
- Add reactions. A one-word "Right." or "Mm-hm." between two long turns does more for realism than anything else you can do by hand.
Generating and exporting
Click Generate segment to voice one section, or Generate episode to do the whole thing. Generation runs on our servers — you can close the tab, and we email you when it is finished.
Export stitches every generated line into one file, plus subtitles. Only lines you have actually generated are included; anything you wrote but never generated is skipped.
Pick MP3 or WAV when you create the episode. MP3 plays everywhere; WAV is a lossless master for Spotify or Apple Podcasts. The choice is locked to the episode, because you cannot mix the two in one file.
What it cannot do yet
Speakers cannot talk over each other. Lines play one after another, so an interruption is a clean cut rather than a real overlap. Writing short reaction lines and em-dash cut-offs ("but that's exactly—") gets most of the way there.
Music beds and intro stings are not available yet either.
Next
- Formats and Speakers — how to configure the brief so the episode sounds the way you want.
- What a Podcast Costs — the points for writing and for voicing.