Tool
Text to speech video generator
Paste any text — your own writing, a script, a story you wrote for a client — and StoryTok returns a finished vertical video: neural narration, captions timed to the word, and a looping background. No Reddit link required and nothing to edit afterwards.
The part that is genuinely hard here is caption timing, and it is the part most tools get wrong. StoryTok does not estimate where words fall from the length of the text; it reads the actual timing marks the speech engine emits while generating the audio. That is why captions stay locked to the syllable even when the narration speeds up over a long sentence — a drift you notice immediately as a viewer and cannot fix without re-timing by hand.
Voice choice changes runtime as well as tone, which matters because billing is per rendered minute. The US voices average roughly 185 effective words per minute; the UK voices and Joey read closer to 140. The same 400-word script is therefore a little over two minutes in one voice and nearly three in another, and StoryTok's estimate accounts for the voice you picked rather than assuming an average.
Because the input is just text, this is the entry point for anything that is not a Reddit post: original fiction, scripted facts, product explainers, motivational scripts, or a series you write yourself and want rendered consistently every day.
How it works
Step 1
Paste your script
Any text works. 300–550 words lands a two-to-three minute video, which is the sweet spot for TikTok monetisation and Shorts retention alike.
Step 2
Pick a voice
Six neural voices across US and UK accents. The estimate updates with your choice, because reading speed varies enough between voices to change the runtime by half a minute on a long script.
Step 3
Style the captions
Word-by-word captions with adjustable colour and position, timed from the speech engine's own marks rather than estimated from text length.
Step 4
Render
About two minutes to a 1080×1920 MP4 with narration, captions and background composited server-side.
Specifics
- Input
- Any pasted text — no Reddit link needed
- Voices
- Joanna, Matthew, Joey (US); Emma, Brian, Amy (UK)
- Caption timing
- Read from real speech marks, not estimated
- Reading speed
- ~185 wpm US voices, ~140 wpm UK voices and Joey
- Output
- 1080×1920 MP4, H.264/AAC, 60fps, no watermark on any paid pack
- Cost
- 3 free minutes at signup, then ~17¢ per rendered minute on the $10 pack
Worth knowing
- Captions are timed from the speech engine's own word marks, so they stay locked even when narration accelerates.
- The credit estimate is voice-aware; a slower voice reserves more, and settlement refunds the difference after the real duration is measured.
- You are billed on the finished video's actual length, not on the estimate shown before rendering.
- No Reddit account, API key or connected service required — text in, video out.
Questions
What voices are available?
Six AWS Polly neural voices: Joanna, Matthew and Joey with American accents, Emma, Brian and Amy with British ones. All are neural rather than the older concatenative voices, which is the difference between narration that sounds synthetic and narration that does not.
How accurate is the caption timing?
It comes from the speech engine's word-level timing marks, generated alongside the audio itself — not estimated from character counts. That is why captions stay aligned through long sentences where narration naturally accelerates.
Does the voice I pick change what I pay?
Indirectly, yes. Billing is per rendered minute and the UK voices read roughly 25% slower than the US ones, so the same script produces a longer video. StoryTok's estimate accounts for the voice you selected, and you are settled on the finished video's real duration.
Turn a script into a video free
3 free trial minutes at signup. No card, no plan, nothing renews.
Try StoryTok freeKeep reading
Product details and pricing current as of August 2026 · Reddit, TikTok, YouTube and Instagram are trademarks of their respective owners; StoryTok is not affiliated with them.