What this is actually good for
If your show is a script you read out loud — a daily news briefing, a solo explainer, a research summary — the single-voice studio handles it directly: paste the script, choose a voice and pacing, generate, download, drop it into your editor.
If the format has more than one voice — a scripted “interview”, a two-host cold open, a dramatized segment — use the multi-speaker studio. Write the script as labelled lines, assign a voice to each speaker, and it renders as one combined file instead of clips you stitch together yourself.
What it isn’t good for: unscripted banter. AI narration reads what you write; it doesn’t riff. If your show’s appeal is two people reacting to each other in the moment, record it live.
Pacing with pause tokens
Podcast narration lives and dies on pacing. Drop pause tokens straight into the script: [[p400]] for a short beat, [[p700]] for a sentence break, [[p1200]] for a segment transition — or any value up to [[p10000]].
Example: “Today’s top story — a shift in interest rates. [[p700]] Here’s what it means for you. [[p400]] First,” — the two pauses do the work a human host does with breath and emphasis.
Which studio fits your format
| Single voice | Multi-speaker |
|---|---|
| Daily news or market briefings | Scripted “interview” segments written as Q&A |
| Solo explainer or commentary episodes | Two-host cold opens |
| Research summaries, monologue shows | Fiction and character-driven audio drama |
From script to MP3
- Paste or import the script. Type directly, or import a PDF or Word file. Long episodes become a few takes.
- Pick voice and pacing. Choose from 322 voices, adjust speed, pitch and volume, add pauses.
- Generate and download. Listen in the browser, optionally add music, then download MP3 or WAV.
One honest note
Listeners can usually tell a script is AI-narrated within the first few sentences, especially on longer shows. That’s not a dealbreaker for briefings, summaries or explainers where the value is the information. But if your show’s appeal is a distinct human voice or chemistry between hosts, text to speech won’t replace that — and pretending otherwise to your audience is a bad idea.