Five recordings made with the same tool this site offers, using the same free voices and the same settings anyone can reach. Nothing was re-recorded, edited or cleaned up afterwards - if a voice stumbles on a word here, it will stumble on it for you too.
Every script was written for these examples. No copyrighted text and no user’s document appears in any of them.
Narration with word-accurate captions
18.2s
A single voice reading straight prose, with the SRT that came out of the same render. The timings are reported by the speech engine as it speaks, so the captions need no adjustment.
Read the script
Every word in this recording has a timestamp, captured by the speech engine at the moment it was spoken. Nothing was estimated from the text, and nothing was recovered by transcribing the audio afterwards. That is why the subtitle file next to this one lines up exactly, without a single manual adjustment.
A short exchange cast from one script, with a different voice per speaker and the gaps between lines set in the editor.
Read the script
— So the whole thing is cast from a single script?
— That's right. You label each line with a speaker, choose a voice for each one, and it comes back as one continuous recording.
— And the captions know who said what?
— Each cue is prefixed with the speaker's name. Useful for interviews, role-play and read-throughs.
The same line read by American, British, Australian, Irish and Indian voices, back to back - the fastest way to hear how far apart they are.
Read the script
The harbour road floods twice a year, so the council finally raised it.
Read in turn by voices from the United States, the United Kingdom, Australia, Ireland and India.
The same sentence in Spanish, French, German, Japanese, Arabic and Hindi. Each is a native voice for its own locale, not an English voice attempting the words.
Read the script
The library closes at eight, so we'll have to hurry.
Spoken in Spanish, French, German, Japanese, Arabic and Hindi by a native voice for each.
One voice at minus 25 per cent, natural pace, and plus 35 per cent - the range that matters for language practice at one end and revision at the other.
Read the script
Slowed down by twenty-five per cent, for a language you are still learning to hear.
At the natural pace of the voice, which is where most listening starts.
And pushed thirty-five per cent faster, for material you have already read once and are revising.