gooseworks-ai/goose-skills / skills/video/create-creator-takes-h3
create-creator-takes-h3
VerifiedGenerate an AI creator talking to camera, saying an approved script, as one continuous track — H3 Max reference-to-video through the GooseWorks fal-proxy (charged to your GooseWorks credits). Plans takes on line boundaries under H3's 15s cap, dry-runs a cost estimate, generates the first take alone so its voice can be locked and passed to every later take, joins takes with measured 0.10s dissolves, and moves each line's timing onto the words actually spoken. Use for any format with a generated creator speaking a script (split-screen, screen inserts, talking-head ads).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 31.6 KB |
| scripts/align_beats.py | script | 3.9 KB |
| scripts/character-prompt.json | config | 2.1 KB |
| scripts/content_review.py | script | 3.3 KB |
| scripts/join_takes.py | script | 12.8 KB |
| scripts/make_character.py | script | 30.0 KB |
| scripts/media_proxy.py | script | 56.8 KB |
| scripts/pair_voice.py | script | 38.1 KB |
| scripts/plan_takes.py | script | 20.7 KB |
| scripts/run_takes.py | script | 19.0 KB |
| scripts/tiers.py | script | 6.1 KB |
| skill.meta.json | config | 278 B |
| tests/smoke-test.md | markdown | 1.3 KB |
| tests/test_action_skin_light.py | script | 7.0 KB |
| tests/test_hands_voice_pace.py | script | 10.1 KB |
| tests/test_make_character.py | script | 4.9 KB |
| tests/test_quality_tier.py | script | 13.5 KB |
| tests/test_run_takes_rejection.py | script | 16.8 KB |
| tests/test_voice_ethnicity.py | script | 34.5 KB |
scripts/align_beats.py
View on GitHub#!/usr/bin/env python3
"""Move every line's start/end onto the words the creator actually speaks. Free, local.
align_beats.py --beats cutlist.json --words creator.words.json --out cutlist.aligned.json
--words is caption-burn's transcribe.py output for the JOINED creator track (reel time):
[{"text", "start", "end"}, ...].
A line's timing comes from the script's estimate until the takes exist, and a generated
voice never speaks at the estimated pace. Unaligned, the reference build cut to its closing
card while the creator was still mid-list, and clipped the last word.
The script's words are matched to the heard words with a sequence match, so one misheard
or dropped word cannot shift everything after it. Each boundary goes midway between the
last word of one line and the first word of the next; the last line ends HOLD seconds after
the final word. Under 70% of the script heard means the takes do not say the script: stop.
The output keeps every other field (footage windows etc.). A window whose slot changed
keeps its `in`; its `out` is re-derived at the same speed, so re-run footage-cutlist
preview.py + cut.py on the aligned list and look at it again.
"""
import argparse
import difflib
import json
import pathlib
import re
HOLD = 0.45
def bare(w):
return re.sub(r"[^a-z0-9]", "", w.lower())
def align(beats, words):
script, owner = [], []
for i, b in enumerate(beats):
for w in re.sub(r"<[^>]*>", " ", b.get("vo", "")).split():
if bare(w):
script.append(bare(w))
owner.append(i)
sm = difflib.SequenceMatcher(None, script, [w for w, _, _ in words], autojunk=False)
t_of = {}
for blk in sm.get_matching_blocks():
for k in range(blk.size):
t_of[blk.a + k] = words[blk.b + k]
first, last = {}, {}
for idx in sorted(t_of):
_, s, e = t_of[idx]
i = owner[idx]
first.setdefault(i, s)
last[i] = e
for i, b in enumerate(beats):
if i == 0:
b["start"] = 0.0
if i + 1 < len(beats) and i in last and (i + 1) in first:
cut = round((last[i] + first[i + 1]) / 2, 2)
b["end"] = cut
beats[i + 1]["start"] = cut
if i == len(beats) - 1 and i in last:
b["end"] = round(last[i] + HOLD, 2)
return len(t_of) / max(len(script), 1)
def main():
ap = argparse.ArgumentParser()
ap.add_argument("--beats", required=True)
ap.add_argument("--words", required=True)
ap.add_argument("--out", required=True)
ap.add_argument("--max-end", type=float, help="clamp the last line to this (the creator track length)")
a = ap.parse_args()
spec = json.loads(pathlib.Path(a.beats).read_text(encoding="utf-8"))
raw = json.loads(pathlib.Path(a.words).read_text(encoding="utf-8"))
words = [(bare(w["text"]), float(w["start"]), float(w["end"])) for w in raw
if bare(w.get("text", "")) and w.get("start") is not None and w.get("end") is not None]
beats = spec["beats"]
before = [(b["start"], b["end"]) for b in beats]
heard = align(beats, words)
print("[align] %.0f%% of script words heard" % (100 * heard))
if heard < 0.7:
raise SystemExit("under 70% of the script was heard; the takes do not say the script")
if a.max_end:
beats[-1]["end"] = round(min(beats[-1]["end"], a.max_end - 0.05), 2)
for (s0, e0), b in zip(before, beats):
if "in" in b and "speed" in b:
b["out"] = round(b["in"] + (b["end"] - b["start"]) * b["speed"], 3)
elif "in" in b and "out" in b:
b["out"] = round(b["in"] + (b["end"] - b["start"]) * ((b["out"] - b["in"]) / max(e0 - s0, 1e-6)), 3)
print(" %-4s %6.2f-%6.2f -> %6.2f-%6.2f" % (b["id"], s0, e0, b["start"], b["end"]))
spec["aligned"] = True
pathlib.Path(a.out).write_text(json.dumps(spec, indent=2, ensure_ascii=False), encoding="utf-8")
print("[align] %s" % a.out)
if __name__ == "__main__":
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |