Narakeet turns a script into a voiceover, and a slide deck into a narrated video. It offers 930 voices in 112 languages and accents. There is no subscription. You buy minutes once and spend them whenever you like.
It comes from Video Puppet Limited, a UK company, and has been running since 2018. This review covers what it does, what a minute costs, and where it shows its age.
The core tool is text to audio. Paste text, or upload a Word document, a plain text file or a subtitle file. Pick a voice and get an MP3, M4A or WAV back. Subtitle files are the clever part. Upload an SRT or WebVTT and the speech lands on the timestamps. You get a dubbed track that already fits the video.
The second tool is video. Upload a PowerPoint, Google Slides or Keynote file with narration in the speaker notes. Narakeet reads the notes, builds the video and adds subtitles. Or write the whole video as a Markdown script, with images, clips and screen recordings dropped in. Change the text, upload again, and you have a new version. No re-recording.
The third tool is speech to text. Upload an audio or video file and get a transcript in more than 100 languages.
Developers get an API and a command line client. Both plug into build pipelines, so documentation videos can rebuild themselves when the product changes.
Narakeet sells minutes of output, not subscriptions. You pay once, the credit sits in your account, and it never expires. Each file you build is charged by its final length in one-second steps.
5 Views
Any paid pack turns the account commercial. Script limits rise to 1,024 KB and uploads to 350 MB. A single transcription can run 60 minutes. It also unlocks the API, SSML markup and batch creation.
Refunds cover unused credit for 14 days. Payment is by card or PayPal. Schools and charities can ask for a discount, but only as an organisation.
Teachers and trainers with slide decks. Speaker notes become narration and the deck becomes a video. When the lesson changes, you edit text, not audio.
Localisers and e-learning teams. Subtitle-to-audio dubbing in 112 languages, without lining anything up by hand.
Developers who ship documentation. Video from a Markdown script, rebuilt on every release through the API.
Anyone who hates subscriptions. Six dollars buys half an hour of audio and stays in the account until you use it.
The free account is a demo. Twenty conversions and a 1 KB script is about 150 words per file. Anything you make with it cannot be used commercially.
Free users wait at busy times. The app has a "we're too busy, come back later" screen, and paid users go first.
The voices are not Narakeet's own. Its FAQ says they come from Amazon, Microsoft and IBM. The same voices exist elsewhere. What you pay for is the workflow around them.
PowerPoint support has gaps. Animations and transitions are dropped, and embedded videos on a slide do not play. Fonts must be embedded or the video looks different from the deck.
Twenty cents a minute is the entry price. Heavy users only reach the cheap rate by paying $500 up front.
Yes, and the free previews make trying it free in every sense. Test the voice on your real script before buying anything.
The subtitle dubbing and the slides-to-video path are the reasons to pick it over a plain text to speech box. If you only want a paragraph read aloud, a cheaper tool will do. If you make training videos and rewrite them often, $45 for 300 minutes is hard to beat.