In short: Turn on Two speakers in Text to Speech, start each line with
A: or B: (or two names such as Host: and Guest:), choose a voice for each speaker and press Generate. You get one MP3 with both voices, up to 10,000 characters per conversion.How to make a two-voice dialogue
- Turn on “Two speakers”It sits right under the text box in Text to Speech. On this page it is already on.
- Write your scriptStart each line with
A:orB:. Names work too:Host:/Guest:,Mia:/Leo:,[Teacher]/[Student]. A line without a name continues the last speaker. - Pick a voice for each speakerTap A, choose a voice in the list, then tap B and choose the second voice. Press ▶ on any voice to preview it first.
- Choose the pause and generatePick Short, Natural or Long pauses between speakers and press Generate Audio. Your MP3 appears in the results list.
What happens behind the scenes
Each voice speaks its own lines. Speechma Pro creates every turn with the voice you chose for that speaker, a few at a time so it stays fast, then joins the clips in your browser:
- Silence trimmed: the empty air at the start and end of each clip is removed, so turns never feel slow.
- Loudness matched: both voices are levelled to the same speech volume, so nobody sounds closer to the mic.
- Natural pauses: a short gap of your choice between speakers, with tiny fades so there are no clicks.
- One file: the result is a single MP3 you can play, download or add to a ZIP with your other clips.
You only once for the whole dialogue. Editing one line and generating again is quick: lines that did not change are reused instead of being made again.
Dialogue scripts you can paste right now
Copy one into the text box, change the details and generate.
- Podcast opener:
Host: Welcome back to Small Business Hour. Today we talk about pricing.Guest: Thanks for having me. Pricing is where most new owners leave money on the table. - Language lesson:
A: Excuse me, how do I get to the train station?B: Go straight for two blocks, then turn left at the bakery. - Customer service demo:
Customer: Hi, my order arrived damaged.Agent: I am sorry to hear that. I can send a replacement today, at no cost. - Explainer video:
Mia: So what does a heat pump actually do?Leo: It moves heat instead of making it. That is why it uses so little power. - Radio ad:
A: Did you see the new Harbor Coffee menu?B: The maple latte? I have had three this week.A: Harbor Coffee. Open early, every day.
Which voices work best together
| Format | Good pairing | Tip |
|---|---|---|
| Interview or podcast | One male and one female voice from the same country | Clear contrast helps listeners follow who is talking |
| Language learning | Two voices of the target language and country | Use the Long pause so learners can repeat |
| Ads and radio spots | A warm voice plus an energetic one | Keep each line under 15 words |
| Stories and kids content | Mix two voices, for example a narrator and a kids or villain character voice | Some character voices read up to 10,000 characters per conversion |
| Training and role-play | Two neutral, clear voices | Name the speakers by role: Manager:, Employee: |
Need more than two speakers or longer scripts?
Each conversion holds up to 10,000 characters of speech. For a long episode, split the script into parts and join the MP3s in any editor. For three or more voices, generate each extra speaker’s lines as a separate clip.
Frequently asked questions
Is the AI dialogue generator included in Speechma Pro?
Yes. It is part of the Text to Speech tool. Commercial use of the audio is included.
How do I tell Speechma Pro which line belongs to which voice?
Start each line with A: or B:, or with two names followed by a colon, like Host: and Guest:. Speechma Pro maps the first name to Speaker An and the second to Speaker B. If no line has a name, lines simply alternate A, B, A, B.
Can I mix a standard voice and an AI voice?
Yes. Each speaker can be any of the 1,900+ voices. A few character voices read up to 10,000 characters, so lines for those voices can have up to 1,000 characters in total.
Will both voices have the same volume?
Yes. Every line is levelled to the same speech loudness before the clips are joined, and silence at the start and end of each clip is trimmed.
Can I use the dialogue in YouTube videos and podcasts?
Yes. You can use the audio in monetized videos, podcasts, ads, courses and client work. Please do not use voices to impersonate real people.
Does it work on my phone?
Yes. The whole dialogue is joined inside your browser on any modern phone, tablet or computer.
The Speechma newsletter
Make better voiceovers, faster
- New voices and features the day they launch
- Script templates and tricks that make AI voices sound human
- Creator stories, community picks and early access