Skip to content

armgit / hypit

Public Imported from GitHub
0

Clone any viral video with AI agents. Not just a script, the whole workflow: swap the face, the words, the B-roll, ship 100 variants in one command, and get your 100M views.

fix: a split shot's parts start from one image, not from each other

rpone4d5fa476cfad
diff --git a/.agents/skills/hypit/references/playbooks/craft/generated-dependencies.md b/.agents/skills/hypit/references/playbooks/craft/generated-dependencies.md
index d911d63d..eb772098 100644
--- a/.agents/skills/hypit/references/playbooks/craft/generated-dependencies.md
+++ b/.agents/skills/hypit/references/playbooks/craft/generated-dependencies.md
@@ -26,17 +26,22 @@ only about the video: what it should say, who is in it, what it is for.
Images therefore generate in rounds: the establishing image is accepted first, and the views derived
from it are generated after. `production-gates.md` Gate 1 is one stage per round, not one stage.
-## A split shot hands over its last frame
-
-A continuous shot that exceeds the generator's duration ceiling is split. The second part opens on
-the **last frame of the first part**, extracted from the accepted take and passed as its first frame.
-
-Every piece of this already exists: a video generation shape that takes a first and last frame, and a
-frame extraction that takes `at="last"`. Only the connection is missing, which is why a split shot
-drifts across its own seam today.
-
-Do not do this across a real cut. A cut is a discontinuity; wiring one would force a match the
-reference never had.
+## A split shot's parts all start from the same image
+
+A continuous shot that exceeds the generator's duration ceiling is split into as many takes as it
+needs. **Every part references the same first-frame image** — the one accepted for that shot — and no
+part references another part.
+
+This is the one place where the "wire what must match" principle does *not* mean chaining. Handing
+part one's last frame to part two would match the seam exactly, and would also make the parts
+strictly sequential: part two cannot start until part one has finished generating and its frame has
+been extracted, so a shot split four ways takes four generations end to end. Pointing every part at
+the same accepted image costs the exact frame match at the seams and buys parallelism: all four
+generate at once, and they still hold the same person, framing, wardrobe and light, because they all
+came from the same picture.
+
+The seam is a small jump in pose and micro-motion, not a change of scene — the same order of
+difference as a cut between two takes of the same setup, which is what the reference has anyway.
## A voice is generated once
diff --git a/.agents/skills/hypit/references/playbooks/craft/visual-continuity.md b/.agents/skills/hypit/references/playbooks/craft/visual-continuity.md
index f76db62e..11a135c4 100644
--- a/.agents/skills/hypit/references/playbooks/craft/visual-continuity.md
+++ b/.agents/skills/hypit/references/playbooks/craft/visual-continuity.md
@@ -5,7 +5,8 @@
- Declare recurring anchors as explicit `media:Image`/`media:Audio` assets or accepted generated
outputs. Reuse the same Artifact edges; do not redraw identity from prose for every take.
- Generate the location once and derive its other views from that image, rather than describing the
- room again per take; a split shot opens on the previous part's last frame.
+ room again per take; every part of a shot split by the duration ceiling references that same
+ accepted image, so the parts generate in parallel rather than in a chain.
`generated-dependencies.md` states both, and the reverse-view geometry below is what decides how
many views a location needs.
- Keep one authoritative reference set and one stable Recipe for a continuous shot group. Split a