Updated
VdoBloom’s AI text to speech tool turns a written script into a spoken voiceover in your browser. Paste your text, choose a voice, and the audio is generated in the cloud in seconds, ready to play and download as an MP3. It works as an AI voice generator for video narration, YouTube and TikTok voiceovers, explainers, ads, e-learning and presentations, with no microphone or editing software.
Is it free? You can try it with the 10 free credits every new account gets, which covers a voiceover of up to 1,000 characters on MiniMax Speech 2.8 or xAI TTS. Pricing is per length: 8 credits per started block of 1,000 characters on MiniMax or 6 on xAI. Playback in the browser works on every plan, and MP3 downloads are included with any paid plan, starting with the Lite pack of 300 credits for $19.50 one-time (credits never expire) or $15 a month.
You get 61 named voices across three engines: 26 voices on MiniMax Speech 2.8, eight expressive voices on Fish Audio S2.1 Pro (80+ languages, detected from your text) and 27 on xAI TTS. MiniMax and xAI have a language menu of 16 languages, including Hindi, Tamil, Spanish, French, Japanese and Arabic. The exact credit cost for your text appears before you click Generate.
Turn any script into a natural voiceover with VdoBloom's Text to Speech. Type or paste up to 10,000 characters, pick a model (MiniMax Speech 2.8, Fish Audio S2.1 Pro or xAI TTS), choose a voice and an optional language,…
Enter the text you want spoken, from a one-line hook to a full narration. The credit cost updates with the length of the text.
Pick MiniMax Speech 2.8 (26 voices), Fish Audio S2.1 Pro (8 voices) or xAI TTS (27 voices) in the Provider menu, then choose a voice from that engine’s list.
For MiniMax and xAI, choose one of 16 languages or leave it on Auto. Setting the language helps with names, accents and scripts that mix languages. Fish Audio has no menu: it detects the language from your text.
Click Generate. The voiceover plays in the browser so you can check it, and paid plans can download it as an MP3 for any video editor.
MiniMax Speech 2.8 offers 26 voices (13 female, 13 male) with different tones; Fish Audio S2.1 Pro adds eight expressive voices in 80+ languages, and xAI TTS 27 more. Switch engines and voices to compare reads of the same line.
English, Hindi and Tamil plus 13 more, from Spanish and Portuguese to Japanese, Korean and Arabic. MiniMax Speech 2.8 itself speaks 40 languages.
Cost is calculated per started block of 1,000 characters and shown before you generate. There is no subscription required and no per-voice surcharge.
The MP3 drops straight into VdoBloom’s video tools or any editor, so a script can become a narrated video in one workspace.
MiniMax and xAI cost 8 and 6 credits per 1,000 characters; Fish Audio is priced per 1,000 bytes of text (about 1.2 credits, rounded up per run), so the choice comes down to voice style, language and script length. MiniMax Speech 2.8 has the largest voice list and broad language coverage. Fish Audio S2.1 Pro covers the most languages (80+) with expressive, natural reads. xAI TTS has a more expressive delivery, which suits hooks, ads and short social clips.
A simple way to decide: write your script once, generate the first two sentences with two or three voices from each engine, and keep the one that fits your brand. A two-sentence test costs 8 credits per voice on MiniMax or 6 on xAI.
| Engine | Voices | Language control | Credits per 1,000 characters | Best for |
|---|---|---|---|---|
| MiniMax Speech 2.8 | 26 | 16-language menu or Auto | 8 | Narration, explainers, multilingual scripts (up to 10,000 characters per generation, split and joined automatically) |
| Fish Audio S2.1 Pro | 8 (Laura, Sarah, Hannah, Selene, Egirl, Adrian, Ethan, Jordan) | Detected from the text (80+ languages) | About 1.2 per 1,000 bytes, rounded up per run (Hindi and CJK use about 3 bytes per character) | Expressive narration in many languages, lowest price on English scripts |
| xAI TTS | 27 (una, leo, eve, ara, sal, rex and 21 more) | 16-language menu or Auto | 6 | Expressive hooks, ads, short social voiceovers |
Text to speech is billed by the length of your script. Every started block of 1,000 characters (letters, spaces and punctuation) costs 8 credits on MiniMax or 6 on xAI, so a 1,001-character script counts as two blocks. Fish Audio is billed per byte of text instead: about 1.2 credits per 1,000 bytes, rounded up for the whole script (an English character is one byte; Hindi, Chinese, Japanese and Korean characters are about three).
With the Lite pack (300 credits for $19.50 one-time, or $15 a month), you can voice 37 blocks of 1,000 characters on MiniMax or 50 on xAI. One-time credits never expire; subscription credits reset each billing cycle. Failed generations are refunded automatically.
| Script | Characters | Blocks billed | Credits (MiniMax / xAI) |
|---|---|---|---|
| Short ad hook or caption | 300 | 1 | 8 / 6 |
| One-paragraph voiceover | 1,000 | 1 | 8 / 6 |
| Product explainer | 2,500 | 3 | 24 / 18 |
| Long narration (one run) | 6,000 | 6 | 48 / 36 |
Voices are listed by name in the tool, and each engine loads its own list when you select it. You can generate the same line with several voices to hear the difference before you commit to a long script.
For MiniMax Speech 2.8 and xAI TTS, the Language menu offers these 16 languages, plus an Auto / Default option. Write your script in the target language and select it in the menu for the most natural pronunciation. Fish Audio S2.1 Pro detects the language from the text and speaks 80+ languages, including English, Hindi, Spanish, French, Japanese and Arabic.
Most scripts on VdoBloom end up inside a video. Common uses include:
Read more: How to add an AI voiceover to any video · AI text to speech guide
The voice reads exactly what you write, so a few edits to the script make a bigger difference than any setting.
Being clear about what the tool does not do saves credits.
Read more: AI Voice Clone
The Language menu lists 16: English, Hindi, Tamil, French, German, Spanish, Italian, Portuguese, Japanese, Korean, Chinese (Mandarin), Arabic, Russian, Indonesian, Turkish and Vietnamese, plus Auto. Telugu, Malayalam, Kannada, Bengali, Marathi, Gujarati, Punjabi and Urdu are not in the menu today.
You can try it free: every new account gets 10 credits, enough for one voiceover of up to 1,000 characters on MiniMax Speech 2.8 (8 credits) or xAI TTS (6 credits). Free accounts can play the result in the browser. Downloading the MP3 requires a paid plan, starting with the Lite pack of 300 credits for $19.50 one-time.
It depends on the script. On VdoBloom, MiniMax Speech 2.8 gives you 26 voices and suits longer narration and multilingual scripts, while xAI TTS has 27 expressive voices that work well for short hooks and ads. MiniMax costs 8 credits per 1,000 characters and xAI costs 6, so generating the same line with a few voices is the quickest way to choose.
There are 61 named voices: 26 on MiniMax Speech 2.8, eight on Fish Audio S2.1 Pro and 27 on xAI TTS (una, leo, eve, ara, sal, rex and 21 more). Fish Audio speaks 80+ languages detected from your text; for MiniMax and xAI the Language menu offers 16 languages, including English, Hindi, Tamil, Spanish, French, German, Japanese, Korean, Chinese (Mandarin) and Arabic, plus an Auto option.
MiniMax costs 8 credits and xAI 6 credits per started block of 1,000 characters. On MiniMax a 300-character hook is 8 credits and a 2,500-character explainer is 24 credits, and the Lite pack’s 300 credits cover 37 MiniMax blocks (50 on xAI). The exact cost is shown before you generate, and failed generations are refunded.
Yes. On any paid plan, the generated speech downloads as an MP3 file that works in CapCut, Premiere, DaVinci Resolve or VdoBloom’s own video tools. Free accounts can listen to the result in the browser before upgrading.
Partly. The Language menu includes Hindi and Tamil; Telugu, Marathi, Bengali, Kannada, Malayalam and other Indian languages are not available for voice. Write the script in Hindi or Tamil and select it in the menu.
MiniMax Speech 2.8 reads up to 10,000 characters in one generation, splitting and joining long text automatically; split anything longer into parts and join the MP3s in your editor. Billing is per started block of 1,000 characters, which makes several shorter runs cost the same as one long run of the same total length.
Spanish is in the Language menu on MiniMax and xAI (Fish Audio detects it from the text) and every MiniMax voice can read it. There is one Spanish setting, without separate Mexico or Spain options. For the most natural Latin American read, write the script in your local Spanish, select Spanish rather than Auto, and test a few voices on the first line. For a clearly regional accent, describe it in AI Voice Design.
There is no emotion slider, so emotion comes from the voice and the script. Test the MiniMax voices Narrator and Storyteller and a few female voices on one page, write short sentences for tension and paragraph breaks for pauses, and keep one voice per chapter. A 10,000-character chapter costs 80 credits on MiniMax.
Yes. Pick one MiniMax voice and use it for every video so the channel sounds consistent. A 9,000-character script (about 1,500 words) costs 72 credits in one run. The Faceless YouTube Video Generator can then plan the b-roll for the narration.
Yes. There is nothing to install: paste text, pick a voice and click Generate in a phone or desktop browser. The credit cost shows before you generate, and the 10 free sign-up credits cover a first voiceover of up to 1,000 characters.
Not on this page. The text to speech tool uses the built-in voices listed above. If you need a voice that does not exist in the list, AI Voice Design creates one from a written description of age, accent, tone and pace.
Under VdoBloom’s terms of service, you retain ownership of the content you generate with the service. Make sure your script itself does not infringe anyone’s rights, and do not use generated speech to impersonate a real person.
Text to speech turns written words into spoken audio. Speech to text (transcription) does the reverse and turns a recording into written text. VdoBloom offers both: this tool for voiceovers and the Audio to Text tool for transcripts.