Skip to main content

What is AI Video?

AI Video lets you turn any clarife document into a fully narrated video — automatically. The AI reads your screenshots and text, writes a narration script, synthesizes speech in your chosen language, and renders the final video with subtitles. The entire pipeline runs in the cloud. You don’t need any video editing software.

How it works

1

Document analysis

AI reads your document content — headings, paragraphs, code blocks, tables, and screenshots — to understand the tutorial flow.
2

Scenario generation

Using Claude (via OpenRouter), AI generates a narration script with timing cues for each step. The script matches your document structure.
3

Text-to-speech

Google Cloud TTS converts the script into natural-sounding audio using SSML markup. 15 languages are supported.
4

Video rendering

Remotion on AWS Lambda composites your screenshots, audio, and subtitles into a final MP4 video. Branding (watermark, intro/outro) is applied based on your plan.
5

Download

When rendering is complete, you get an MP4 video file and an SRT subtitle file ready for download.

Supported languages

AI Video supports 15 languages for text-to-speech narration:
LanguageCodeLanguageCode
EnglishenPolishpl
GermandeFrenchfr
SpanishesItalianit
PortugueseptDutchnl
SwedishsvNorwegiannb
DanishdaFinnishfi
CzechcsRomanianro
Turkishtr

Requirements

  • A clarife document with at least one screenshot block
  • AI credits (3 free during trial, then purchased in packs)
  • An internet connection (rendering happens in the cloud)
AI Video requires credits. No plan includes credits by default. You receive 3 free generations during the 14-day trial. After that, purchase $7 credit packs (20 generations each) from the Billing page.

Output format

PropertyValue
Video formatMP4 (H.264)
SubtitlesSRT file
ResolutionMatches your screenshot dimensions
AudioSynthesized TTS narration
DurationVaries by document length (typically 1-5 minutes)

Typically 2-5 minutes depending on document length and the number of screenshots. You can close the editor and come back — the generation continues in the background.
The generated video is a final MP4 file. If you want changes, edit the source document and regenerate. You can also download the SRT subtitle file separately.
AI Video requires at least one screenshot block in your document. Text-only documents cannot be converted to video.
There is no hard limit, but very long documents (50+ blocks) may produce longer videos and use more rendering resources. For best results, keep documents focused.