Studio key
Enter the key the installer printed on your server. It is saved on this phone only.
Clean a take
Pick a video you recorded. Pauses, hesitation sounds and false starts are removed on your server, and the finished take waits in your Library.
Cut hesitation soundsum, uh, hmm, अं, हम्म…
Cut false starts and stutters“Knee pain is— knee pain is common” keeps only the second try.
Smart zoom on cutsA 6% punch-in alternates at each cut so the edit reads as intentional.
Balance loudnessEvens the volume to social-media level and stops clipping.
Reduce background hissLight noise reduction. Leave off for clean rooms.
Advanced
Also cut discourse wordslike, basically, okay, मतलब, वो, तो, bueno, pues, يعني… These can change meaning — check the result before publishing.
You can leave this page — the take keeps processing on the server and appears in your Library.
Clean take ready
pauses
hesitations
false starts
Make a Twin video
Type the script. Your cloned voice speaks it and your lips are synced on a base clip — a finished reel without recording.
Auto reads each sentence: Devanagari is spoken as Hindi, Arabic as Arabic, Latin as English or Spanish.
Studio polishGentle denoise, sharpen and colour — a cleaner, crisper face.
Burn captions into the videoSentence captions in the spoken script. A separate .srt is always made.
Voice onlyJust the cloned voice track (for slides and voice-overs). No base clip needed.
Renders take a few minutes. You can leave — the video appears in your Library when it is ready.
Twin video ready
length
sentences
languages
Library
Every take stays on your server until you delete it. Open this page from any phone or the Mac.
Twin assets
Used by the Twin (text → your voice → your face). Add them now so they are ready when the Twin goes live.
Voice samples
60–90 seconds of clean speech, half English and half Hindi, no music.
Base clips
30–60 seconds of you in frame, eyes on the lens, natural nodding, no talking. Same light and clothes for every clip.
Settings
Engine
Version—
Speech analysis—
Takes in queue—
Address—
Render
Voice—
Lips—
Workers—
