한국어 
日本語 
English 
Type a script.
Get a finished training video.
Paste a draft or upload a PDF. ReFlash writes the narration, generates the images, times the subtitles, and hands you an editable timeline. One at a time, or a hundred at once.
per training video
long doc → split output
ZIP bundle export
Two ways to generate
-
Paste a script or drop in a PDF or DOCX. The AI splits it into scenes and, for each scene, writes narration, generates images, and times the subtitles in parallel. Then you take over on an NLE timeline — reorder scenes, swap images, fix captions, adjust speed, set transitions.
Split → TTS → images → subtitles → timeline -
Fan out a long document into many videos in parallel. ReFlash branches one source into separate projects and batch-renders a hundred 3–5 minute videos in roughly an hour. Finished files land in a single ZIP. Built for course catalogs, onboarding tracks, and chaptered product manuals.
100 videos / hour · ZIP bundle export
Four things the AI builds
Voice, image, subtitle, transition — assembled per scene, no editor required.
AI narration
auto-dubbed
TTS reads each scene out loud. Tune speed from 1.0x to 2.0x to land on the runtime you want.
Per-scene images
consistent style
The first scene anchors a style, and the rest follow. Regenerate only the frames that miss.
Auto subtitles
with style presets
Captions follow the narration timing. Pick a preset for font, size, and color — burned straight into the video.
One-click export
MP4 · ZIP
Pick resolution and codec, get an MP4. Bulk runs come back as a single ZIP.
