AI clone recipe: the files behind https://learnwithhasan.com/guide/ai-clone-of-yourself/

prompts/
  talking-head-with-video-reference.txt       the short prompt (Seedance 2.5, with a video reference)
  talking-head-long-version-for-comparison.txt the long first version, for comparison only
  talking-head-no-video-reference-timed-beats.txt  a talking head with timed beats, no video reference
  scene-food-hall.txt                          the food hall (the scorpion)
  scene-cafe.txt                               the cafe
  scene-street.txt                             the street

  These are the exact prompts I ran, copied byte for byte from the files that made the
  renders. Swap "he" for "she" and the quoted line for your own line. @Image1 to @Image3
  (or @Image4) are your registered photos, @Video1 your video reference, @Audio1 your
  recording, in the order you attach them.

relink_voice.py            puts your real recording back on a render (run it with --help)
modelark_seedance.py       submits a Seedance 2.5 render on ModelArk, waits, downloads
requirements.txt           pip install -r requirements.txt  (plus ffmpeg on your PATH)

RECORDING.md               how to record the line
PHOTOS-AND-VIDEO-REFERENCE.md  the photos and the 14 second video reference
HEYGEN-FOOTAGE.md          the footage that fixed my HeyGen twin, and the two-camera trick
TESTING.md                 what each test costs, and the rules that saved me money

Prices are what I paid in September 2026. Check the pricing pages before you plan a batch.
