From written script to multi-host podcast audio
Breeze Blue performs the podcast script you already wrote. Give every host and guest a distinct AI voice, direct each line’s pacing and tone in plain language, and export finished audio into your usual edit.
A distinct voice for every host and guest
Podcasts live or die by voices listeners can tell apart. Build your cast from three sources — pick from the library, design a voice from a description, or clone your own.
Pick from the Voice Library
Browse a curated library where every voice has a personality of its own and a playable preview. Pair a warm, unhurried host with a sharper, quicker co-host so listeners always know who’s talking.
Explore the public voice libraryDesign a co-host from a description
Describe the voice your show is missing — the age, the attitude, the energy — and audition generated candidates in seconds. When one fits, save it and it becomes a reusable member of your cast.
Design a podcast voice from a descriptionClone your own hosting voice
Turn a short authorized recording of your own voice — MP3 or WAV, at least 5 seconds — into a reusable cloned voice, with an explicit consent step. We analyze the first 60 seconds and keep the best 30 for cloning. Script an episode, and your voice reads it even when you have no time to record.
See how authorized voice cloning worksDirect every line like a showrunner
The voice decides who is speaking; direction decides how the line lands. Write instructions in plain language — for the whole episode or a single line — and the voice performs to match: pacing, tone, hesitation, energy.
From a scripted episode
Welcome back to the show. Today’s story starts with a phone call nobody expected.
Honestly… I wasn’t sure I should even talk about this publicly.
Let’s start at the beginning. What happened that night?
On our public voice-direction benchmark (2026-08-17), directed reads score 4.25 overall Voice Direction Score while keeping 0.67 speaker similarity — the delivery changes, but the host still sounds like the host.
Built for full episodes, not one-liners
A 40-minute episode is a long script. Produce it in segments, fix single lines without re-rendering everything, and keep every voice consistent from cold open to outro.
Paste your script
Bring the episode script you already wrote — intro, segments, ad reads, outro. Split it by speaker and section, so each block carries its own voice and direction.
Assign voices to speakers
Map each host and guest to a voice from your cast. A saved voice performs the same in episode one and episode one hundred, so your show keeps a stable sound.
Generate section by section
Work through the script in segments. Listen, adjust the instruction on a line that doesn’t land, and regenerate just that line — no need to redo the whole episode.
Export and edit as usual
Download the finished audio and assemble it in your editor with your music, ad spots, and mixing — then publish through your regular hosting workflow.
What Breeze Blue does — and doesn’t
Breeze Blue is not an end-to-end podcast maker. It doesn’t plan episodes, record live conversations, host your feed, or publish to platforms. It does one thing well: performing your written script as directed voice audio that you take into your normal editing and publishing workflow.
Export and use your episodes commercially
The output is standard audio — ready for any editor, any workflow, any host.
Clean files for the edit
Audio generated in Text to Speech and Studio downloads as WAV, so your edit starts from clean, uncompressed audio; voice previews from the library download as MP3.
Commercial use
Audio generated on paid plans can be used in commercial projects — including monetized podcasts, branded shows, and the ad reads inside your episodes.
Free to start
Start free and claim free credits each day, then move to a paid plan with more monthly credits as your publishing schedule grows.
Frequently asked questions
Does Breeze Blue plan, record, host, or publish my podcast?
No — and that’s deliberate. Breeze Blue turns the script you wrote into performed voice audio. Episode planning, live recording, feed hosting, and distribution stay in the tools you already use; you drop the exported audio into that workflow.
Can each host and guest have a different voice?
Yes. Assign a separate voice to every speaker in your script — picked from the voice library, designed from a text description, or cloned from your own authorized recording. Saved voices are reusable, so your cast stays consistent across episodes.
Can I clone my own voice for my show?
Yes, from a short authorized sample: Web and API accept MP3 or WAV of at least 5 seconds, with an explicit consent statement. We analyze the first 60 seconds and keep the best 30 for cloning. You can only clone voices you have the rights to use.
How do I control pacing, tone, and emotion line by line?
Write the direction next to the line in plain language — “slower, almost a whisper, like sharing a secret” — and the voice performs to match. For single moments, short bracket cues like [sigh] or [giggle] shape the delivery inside a line.
How does it handle a full-length episode script?
Produce long scripts in segments: generate a section, listen, and regenerate individual lines that need a different read. Because each speaker’s voice is a saved asset, a 40-minute episode keeps consistent voices from the first segment to the last.
Can I use the audio in a monetized podcast?
Audio generated on paid plans can be used commercially, including monetized and sponsored shows. See the details for each plan on our pricing page.
What audio do I get, and how do I bring it into my editor?
Generations from Text to Speech and Studio download as WAV, so you can drop them straight into any editor — add music, ad spots, and mixing there, then publish through your usual host.
Is there a free way to try podcast voices?
Yes. The demo on this page lets you direct a voice without an account, and signing up is free with claimable daily credits. Paid plans add more monthly credits and voice slots as your production scales.