Four inputs. About six minutes. Your clone goes operational the moment you hit deploy.
// total training time
about six minutes, total.
Most of it is the audio recording. The rest is paperwork your double does for you the first time you sign in.
STEP 01~ 90s✓
Face
STEP 02~ 3m●
Voice
STEP 03~ 2m○
Cadence
STEP 04~ 30s○
Deploy
Step 01 — Face ● ACTIVE
// identity anchor
One reference photograph anchors your clone to your bone structure. Higher resolution = more durable identity. Optionally add a full-body shot to unlock body-aware generation.
＋
Face referenceDrop a JPG/PNG, or click to upload
RECOMMENDED · 1080×1350+ · 1.0 MB+
＋
Body shot / optionalUnlocks full-body scenes (+10% match)
FULL-LENGTH · NEUTRAL POSE
Step 02 — Voice ● ACTIVE
// timbre + cadence + breath
Read the paragraph below into your microphone, or upload a 60-second audio clip (podcast, voice memo, YouTube cut). The double captures timbre, pace, vowel shapes, and how you breathe between thoughts.
// SCRIPT · ~58 SECONDS
"Hi — my name is what I sign with. I'm here building a clone of myself because I'd rather be making things than performing them. The double speaks the way I speak. It breathes the way I breathe. It pauses the way I pause. When you hear it later, on a livestream or in a podcast, you should not be able to tell the difference. That's the test."
RECORD
live
♪
Or upload audio60s+ of you speaking — podcast cut, voice memo, lecture
MP3 · WAV · M4A · 60s+
~38s training time
Step 03 — Cadence ○ READY
// the way you turn a phrase
Paste 10 of your own sentences — DMs, captions, voice notes transcribed. Or upload a text/PDF archive. The double learns vocabulary, cadence, opening moves, the punctuation you favor.
¶
Or upload an archiveSubstack export, Notes folder, transcripts, anything in your voice
TXT · MD · PDF · DOCX · CSV
Step 04 — Deploy ○ PENDING
// connect platforms
Connect the platforms you want your clone to operate. We never post without your approval until you flip autopilot on.