How Text to Speech works: voices, limits and price

Text to Speech turns written text into a spoken audio file. People use it for a voice-over for slides or a video, to listen to an article instead of reading it, to hear how a foreign phrase is pronounced, or to send a message in their own cloned voice. It works on web.vocalingo.app, iOS and Android.

What text you can enter

One run takes up to 10,000 characters, about 11 minutes of speech at a normal pace. A counter under the field shows how many characters you have.

• Type the text or tap Paste to insert it from the clipboard.

• Dictate it with the microphone button: the app turns your speech into text.

• Send it from another tool: after Speech to Text or Video to Text, tap Text to speech under What next?.

If the text is longer, the app shows "Text is too long. Maximum 10000 characters." Split it into parts and voice them one by one.

How to voice a text

1. On the Tools tab, open Text to Speech.

2. On the Text tab, enter the text and tap Continue.

3. On the Voice tab, choose a voice. The play button next to a voice plays a sample. By default the MiniMax voice Expressive Narrator is selected.

4. If needed, set up the voice below the list: instructions for OpenAI voices, emotion, speed and pitch for MiniMax, stability, similarity and style for ElevenLabs. Tap Continue.

5. On the Result tab, check the estimated cost and tap Generate.

While the audio is being made, the button shows "Generating audio..." with a percentage, and Cancel appears below it.

Which voices you can choose

The voice picker has seven tabs: Favorites, Recent, My voices, MiniMax, Qwen 3, OpenAI and ElevenLabs. MiniMax and ElevenLabs sound the most natural in long texts, OpenAI voices can be steered with instructions, and My voices holds the voices cloned from your recordings. A comparison of all groups with prices: which Text to Speech voice to choose. How to search, preview and save voices to Favorites: choosing a voice.

How to add laughter, pauses and the right stress

With MiniMax and ElevenLabs voices, the Text tab shows tags. Tap a tag to copy it and paste it into the text where it should sound.

• MiniMax: (laughs), (sighs), (coughs), (crying), (gasps), (groans), and <#1#> for a one-second pause.

• ElevenLabs: [excited], [sad], [angry], [curious], [whispers], [laughs], [sighs], [crying], and an ellipsis for a pause. Too many tags in a row can make the sound glitch.

• If a word is stressed on the wrong syllable, write the stressed part in capital letters, for example REcord (an album) and reCORD (to record).

What you get and where to find it

The Result tab shows a player with the audio. Next to the player there is one button: in the app it opens the share menu, so you can send the file or save it on the device; on the web it downloads the file. To change something, edit the text or the voice and tap Generate again.

Every result goes to History: tap the clock icon at the top. Without Premium, the last 3 results are kept, with Premium all of them. Details: History: saving, restoring and deleting results.

How much it costs

The price depends on the voice and the length of the text. Tokens are charged for a finished file: if generation fails, nothing is charged or the tokens come back.

The minimum for any run is 1 token. Instructions, emotion and other settings don't change the price. Prices for other tools: token pricing for each tool.

What Text to Speech can't do

• It doesn't translate. The voice reads the text in the language it's written in. To get speech in another language, translate the text first in Text Translation.

• It can't change the voice in an existing recording. For that there is Voice Changer.

• The result has no What next? block, so the audio isn't passed on to other tools automatically.

If generation takes too long, the app shows "Voice generation took too long. Please try again." Tap Try again, and if it fails again, pick another voice.