How to Convert Text to MP3 with an AI Voice

To turn text into an MP3, open the studio, paste your text into the Text to Speech tab, pick a voice and press Generate. You get an MP3 file of your text read aloud in that voice, which you can play in the page and download.

Use it in your browser; the audio is generated in the cloud (Standard on our own servers, HD by a third-party voice provider), so there's nothing to install and you don't need a graphics card. You sign in with Google, and new accounts get 20 free credits.

What you get

• An MP3 file of your text in the voice you picked, playable in the page.
• A Download MP3 button that saves the file to your device.
• A copy in your recent history to play or download again for 30 days, or 6 months once you've made a purchase.
• No watermark on the audio.

How to convert text to MP3, step by step

1. Open the studio and sign in with Google.
2. Choose the Text to Speech tab.
3. Paste or type your text into the text box.
4. Under Quality, choose Standard or HD.
5. Pick a voice. Press Preview to hear a short clip of it first.
6. With Standard, choose the language your text is written in. With HD, the language is detected from your text, and you can set the speed from 0.7× to 1.2×.
7. Check the cost and your credit balance shown in the studio.
8. Press Generate. If other jobs are queued, the studio shows how many are ahead of yours.
9. Play the result, then press Download MP3.

Standard or HD voices

Standard

• 23 preset voices, each speaking all 7 Standard languages: English, French, German, Spanish, Italian, Portuguese and Dutch. You pick the language from a dropdown.
• Cost: 1 credit per 200 characters, rounded up, with a minimum of 1 credit.
• Limit: up to 3,000 characters per request.
• Where it runs: on our own servers, on Pocket TTS, an open-source speech model made by Kyutai and released under CC-BY-4.0. Your text isn't sent to an outside voice provider.
• Speed: on our server, a 14-second line took about 6.5 seconds to generate when nothing else was queued, about 2x faster than real time.

HD

• 16 premium voices from a commercial cloud voice model.
• 29 languages, detected automatically from your text: English, Chinese, Japanese, Korean, Hindi, German, French, Spanish, Portuguese, Italian, Indonesian, Dutch, Turkish, Filipino, Polish, Swedish, Bulgarian, Romanian, Arabic, Czech, Greek, Finnish, Croatian, Malay, Slovak, Danish, Tamil, Ukrainian and Russian.
• Speed control from 0.7× to 1.2×.
• Cost: 3 credits per 100 characters, rounded up.
• Limit: up to 5,000 characters per request.
• Where it runs: HD text is sent to a third-party voice provider, which generates the audio.

A 1,000-character script costs 5 credits in Standard and 30 credits in HD. Choose Standard when your text is in one of its 7 languages. Choose HD for languages such as Chinese, Japanese, Hindi or Arabic, for speed control, or for requests longer than 3,000 characters.

Length limits and long scripts

Each press of Generate is one request, with its own limit: up to 3,000 characters in Standard and up to 5,000 in HD. Credits are counted per request and rounded up, so in Standard any request of up to 200 characters costs 1 credit and a full 3,000-character request costs 15. In HD, a 100-character line costs 3 credits.

For a longer script, split it at paragraph or sentence breaks into parts that fit the limit, and generate each part with the same voice, language and speed so the parts sound alike. Download the MP3 files, number them in order, and join them in any audio or video editor.

Free credits and what they cover

When you sign in with Google for the first time, you get 20 credits, valid for 365 days. They're enough for 4,000 characters of Standard speech or about 660 characters of HD speech.

You can spread those credits across several requests, but each request still has its own limit. In Standard, 4,000 characters means, for example, one request of 3,000 characters and another of 1,000, not a single 4,000-character request.

The same balance pays for Voice Clone and Speech to Text. Daily check-ins add more credits, up to 20 for a full 7-day streak, and those last 30 days. For more, see the monthly plans and one-time packs on pricing: plan credits are valid until the end of each billing month and don't roll over, while pack credits never expire. If a generation fails, its credits go back to your balance automatically.

Commercial use

You can use the MP3s you make here commercially, with Standard and HD voices alike, for example in monetized YouTube or TikTok videos, ads, client work, paid courses and podcasts. There's no watermark on the audio. If you use Voice Clone, only clone your own voice or a voice whose owner has agreed.

Questions people ask

Is text to MP3 free here?

You can start for free: the 20 welcome credits cover 4,000 characters of Standard speech. After that, requests use credits from a plan or a pack, and the cost is shown before you generate.

How long can the text be?

Up to 3,000 characters per request in Standard and up to 5,000 in HD. Split longer scripts and join the parts after download.

Which languages can I convert?

Standard reads 7 languages and HD reads 29, including Chinese, Japanese, Korean, Hindi and Arabic. To hear your own voice in one of 21 languages, see the voice cloning guide.

Can I choose the voice?

Yes. The Standard voices are Alba, Anna, Vera, Fantine, Charles, Paul, Eponine, Azelma, George, Mary, Jane, Michael, Eve, Marius, Javert, Bill, Peter, Stuart, Caro, Estelle, Giovanni, Lola and Daan. The HD voices are James, Arabella, Mark, Adeline, Liam, Callum, Laura, Brian, Bella, Aria, Lucy, Emma, Edward, Finn, Hope and Sam. Every voice has a short preview clip you can play in the voices section.

Can I get a transcript or subtitles from audio?

Yes, with the Speech to Text tab in the same studio. The audio to SRT guide shows how.