50% off yearly plansLimited time — ends in 50% off yearly plans · ends in

Multi-Speaker Text to Speech

Write a dialogue with A: and B: lines, pick two voices, and get the whole scene in one audio file.

Made with AdAnt“A: Did you finally fix the sink? B: [proudly] Fixed it, cleaned it…” · Sarah and Ethan · Fish Audio Drama 3

  • Sarah and Ethan (ad skit)

    “A: Did you finally fix the sink? B: [proudly] Fixed it, cleaned it, and now it smells like lemons. A: [suspicious] You bought the spray, didn't you? B: [grinning] Best twelve dollars I ever spent.”

  • Adrian and Hannah (audio drama)

    “A: [low and urgent] Don't open that door. B: [scared] Why? What's behind it? A: [quietly] I don't know. That's why.”

  • Hannah and Ethan (podcast)

    “A: So what's the one thing nobody tells you about starting a garden? B: [laughing] That the snails always win. Every single time. A: [sighing] I was afraid you'd say that.”

  1. Write the dialogue, one line per speaker, starting with A: or B:.
  2. Pick a pair of voices and press Create.
  3. Download the scene, or make a video with it in AdAnt Studio.

Multi-Speaker Text to Speech: Two Voices, One Audio File

Multi-speaker text to speech reads a dialogue in more than one voice in a single run. Start each line with A: or B:, pick a pair of voices, and Fish Audio Drama 3 acts the whole scene, following directions like [laughing] in brackets. You get one audio file for an ad, a story or a video.

Ad skits

Two people, one problem, one product, in fifteen seconds.

Audio drama

Voice a scene with a cast instead of one narrator.

Podcast-style clips

Turn a written exchange into a short back-and-forth.

Multi-Speaker Text to Speech FAQ

Tips for a better result
  • Keep each turn to one or two sentences.
  • Put a direction in brackets at the start of a turn: [whispering], [laughing].
  • Use A: and B: only. Each pair has two voices.
  • If a turn comes out in the wrong voice, press Create again.
Is it free?

This tool uses paid credits; creating an account is free. Runs use credits from a plan: a 1,000-character scene takes about a twentieth of a credit.

Does every turn get the right voice?

Not always. In our tests, about two takes in three kept every turn in its own voice. If one slips, press Create again.

Which models can voice a dialogue?

Fish Audio Drama 3, which this page uses. On AdAnt, Eleven v4 and MiniMax Speech read one voice per run (as of November 2026).

Can I use my own voice?

No. The voices are stock voices from Fish Audio's library, and AdAnt doesn't clone voices.

More AI tools

See every AI tool

Make a video with this voice

Put the voice on an AI avatar or your product video in AdAnt Studio.

Open AdAnt Studio