Official BytePlus partnerBytePlus, a ByteDance company
AI podcast voices with two-speaker dialogue
Write a conversation for two hosts, give each speaker a voice and the studio generates every line in order. A dialogue holds up to 12 turns of up to 600 characters each, cast from 45 voices in 14 languages.
- Two speakers
- Up to 12 turns
- 600 characters per turn
- 2 credits per 1,000 characters
Last updated
On this page
How two-speaker mode works
- 1
Switch to Two speakers
Open Text to Speech and switch the speaker mode to "Two speakers".
- 2
Cast Speaker A and Speaker B
Pick a voice for each host from the library, including your own cloned voice under My Voices.
- 3
Write the turns
Add up to 12 turns of up to 600 characters each and set who speaks every turn.
- 4
Generate and play all
Every turn is generated in its speaker's voice. Play all plays the conversation in order, and each line has its own player and download.
Conversational voices to cast
One conversational voice per language, under its real studio name. Mix any two voices: the hosts do not have to share a language.
What creators make with it
- Podcast cold opens, intros and trailers
- Interview-style segments with a host and a guest
- Two-host explainers and news recaps
- Ads with two characters
- Language-learning dialogues
Longer episodes
One dialogue holds up to 12 turns. For a full episode, write it in segments, generate each segment and join the downloaded lines in your audio editor, where you can also add music and a sting.
For a scene you would rather describe than script line by line, Audio Generation can voice two speakers from one prompt with "Speaker A:" and "Speaker B:" lines. Its voices come from the prompt, so they can change between runs; Text to Speech keeps the same two voices every time.
Pricing
Each turn is billed as its own clip at 2 credits per 1,000 characters, with a minimum of 2 credits per turn, so a full dialogue of 12 turns of 600 characters costs 24 credits. The studio shows the total before you generate, and failed clips are refunded automatically.
Text to speech by language
One page per language with at least two preset voices: its voices, tips and common questions. Also in Text to Speech, with one voice each: Italian (Matteo), Vietnamese (Mai) and Thai (Napha).
AI podcast voice questions
How many speakers can a dialogue have?
How long can a dialogue be?
Can I host with my own voice?
Can the two hosts speak different languages?
What does a dialogue cost?
Which platform does the audio studio run on?
Write your first two-host scene
Switch to Two speakers, cast two of 45 voices and generate up to 12 turns.
Keep exploring
- AI voice generator
- AI text to speech
- AI voice cloning
- Multilingual text to speech
- AI voice translator
- AI voiceover generator
- AI audiobook narrator
- Add a voiceover to an AI video
- Clone your voice safely
- Timestamped audio prompts
- Two-speaker AI dialogue
- AI audio studio
- Add a voiceover to an AI video
- Seedance 2.0 video generator (native audio)
- Seedance 2.5 video generator (native audio)
- Guides
- Acceptable Use Policy