Write the script, pick a voice, get a voiceover
This text to speech tool reads your script aloud in one of 30 AI voices, or as a back-and-forth between several speakers with a voice each. You get an MP3, WAV or OGG file in your media library, ready for a podcast intro, an ad spot or a video voiceover.
How it works
- 1
Write or pick a sample
Type your own text, or start from a sample like a podcast intro or an ad spot.
- 2
Choose the voice
Single voice or multi-speaker. Hit ▶ to hear a sample, then set format and language.
- 3
Generate
It runs in the background and the file lands in your media library.
What you can make
Podcasters
A consistent intro and outro for every episode, without recording them again each week.
Newsletter writers
Offer a read-aloud version of each issue for subscribers who would rather listen.
Small business owners
Voice a short ad spot or a promo video yourself, in Indonesian or English, without booking a studio.
Storytellers and educators
Script a dialogue between two characters, or a teacher and a student, and give each one a different voice.
This text to speech tool reads your script aloud in one of 30 AI voices, or as a back-and-forth between several speakers with a voice each.
Generate your first voiceoverSpecs & limits
| Text length | 1 to 5,000 characters, speaker labels included |
|---|---|
| Voices | 30, each with a playable sample. Kore is the default |
| Modes | Single voice, or multi-speaker with a voice and label per speaker |
| Languages | Auto-detect, or pick from 17 including Indonesian, English (US/UK), Japanese and Arabic |
| Output | MP3 (default), WAV or OGG (Opus) |
| Fine-tuning | Style instructions up to 500 characters, temperature 0 to 2 |
| Billing | Per 100 characters, rounded up, with a live estimate |
Tips for better results
- Play a few samples before you commit. Kore is firm, Puck is upbeat, Charon is calm, and the same script lands very differently in each.
- In multi-speaker mode, start every line with the speaker label, like "Host:" or "Guest:". A line without a label has no voice assigned to it.
- Use the style instructions under "More" for delivery, not content. Something like "warm and slow, like a bedtime story" works better than rewriting the script.
- Set the language yourself when the text mixes two. Auto-detect picks one, and an Indonesian script full of English terms can come out with the wrong accent.
- Not quite right? Use "Re-run this prompt". The next take is saved as a new version, so you can compare it with the first one.
Frequently asked questions
Can I make a conversation with more than one voice?
Yes. Switch to multi-speaker, add several speakers, and give each one a voice and a short label (up to 32 characters). Then write each line starting with that label, for example "Host: Welcome back."
Does it speak Indonesian?
Yes. Indonesian is one of 17 languages you can pick, and auto-detect handles it too. Pick it explicitly if your script mixes in a lot of English.
How long can the text be?
Up to 5,000 characters per run. In multi-speaker mode the speaker labels count toward that total, so a long dialogue fills it a little faster.
Is it free?
No. Each run spends AI credits from your Wujud account, billed per 100 characters of text, and the dialog shows the estimate before you generate. Credits are bought as a pack and new accounts start with none. The tool needs a paid plan.
Can I use the audio commercially?
Yes, including ads, client videos and paid podcasts.
Related tools
An animated waveform and captions over your podcast clip, so it still lands with the sound off. Square, vertical or landscape. Make your first one now.
B-roll, voiceover and captions from a single topic. 10 to 60 seconds, and nobody ever has to appear on camera.
Got a new voice track for a video you already shot? Paste the audio link and the speaker’s mouth is re-synced to match it, no reshoot. Sync your first clip.