render-narrated-ugc-wardrobe-stitch

Name: render-narrated-ugc-wardrobe-stitch
Author: gooseworks-ai

1.0kstars193forks

How to install render-narrated-ugc-wardrobe-stitch?

gooseworks-ai/goose-skills/render-narrated-ugc-wardrobe-stitch

$npx -y skills add gooseworks-ai/goose-skills --skill render-narrated-ugc-wardrobe-stitch

Installs into the current project.

Summary

Assemble a narrated-UGC "stitch reply" ad from a config — a single spoken VO carries a verbatim testimonial while ~30 per-cut i2v clips (one creator across ~5 wardrobes in ~3 worlds, plus product B-roll) are each trimmed to their EDL window built from the VO's Whisper word bounda

Files · 1

View on GitHub

SKILL.md

1	# render-narrated-ugc-wardrobe-stitch
2
3	Assemble a narrated-UGC "stitch reply" ad from a config: a fast-cut vertical testimonial
4	where a single spoken VO carries a verbatim ~13-sentence reversal-hook monologue over ONE creator
5	across ~5 wardrobe changes in ~3 micro-worlds, interspersed with product B-roll (capsule macro,
6	unboxing, a landing-page scroll), ~30 hard cuts on the VO cadence, closing on a brand end card.
7	This capability is the FREE, deterministic assembly — trim-to-EDL, hard-concat, the VO+music
8	mix, the karaoke-pop caption burn, the landing-page zoompan, and the end-card append.
9
10	`scripts/config.example.json` is the worked example (Bioma "Do NOT buy Bioma Probiotics", ~37s
11	1080×1920 9:16, ~30 body cuts + a ~2s end card); `scripts/PIPELINE.md` maps every config block to
12	its source step and `scripts/README.md` documents the free assembly.
13
14	## Run
15
16	This is the FREE, deterministic assembly stage — it spends nothing. The paid inputs are
17	separate capabilities — the spoken VO (`create-vo-elevenlabs`) Whisper-aligned so the WORD
18	BOUNDARIES set the cut grid; one locked creator (`create-image-gpt-image-fal` anchor + ~5 wardrobe
19	edits chained off the anchor) + 3 world wides + per-cut start-frames (`create-image-fal` product
20	composites); and one Veo/Seedance i2v clip per cut (`create-video-fal`). Given the VO +
21	`vo-final.words.json` + `edl.json` + one clip per cut + a Playwright landing-page PNG + the brand
22	end-card PNG, `render-narrated-ugc-wardrobe-stitch` trims each clip to its EDL window, hard-concats
23	on the VO cadence, mixes the VO over the ducked bed, burns the karaoke-pop captions, appends the
24	end card → the master. Re-cuts reuse the existing VO / start-frames / clips and cost $0.
25
26	## Contract (the free assembly)
27
28	- The spoken VO carries the narrative — lock it FIRST. The VO IS the narration bed; the whole
29	ad is cut to it. Never plan the cut grid before the VO is locked and Whisper-aligned.
30	- Build the EDL from the VO's Whisper word boundaries. ~30 role-tagged cuts (`hook`, `feature`,
31	`reaction-insert`, `payoff-hold`, `b-roll-insert`, `landing-page`); snap every cut window to the
32	word boundaries. The payoff line gets a HELD `payoff-hold` beat (~3× mean shot length).
33	- Hard cuts via `filter_complex concat`, not the demuxer. Trim each clip to its EDL window and
34	hard-concat with `filter_complex concat` — the `-f concat` demuxer drops the audio when a
35	drawtext/scale step shaves a clip a few ms below its window. No dissolves.
36	- Karaoke-pop captions on every word, throughout. From the VO's `vo-final.words.json` (VEED
37	Whisper preset, bold yellow), on every word; re-spell brand tokens Whisper mishears against the
38	locked script ("synbiotic" over "symbiotic"; keep "I'ma" verbatim) — never edit the script to
39	match Whisper. Captions are suppressed over the end card. If VEED mis-captions a brand token,
40	hand-patch that sentence with local ASS karaoke.
41	- Product B-roll breaks up the talking head. Capsule macro, unboxing, and a landing-page
42	scroll are interspersed with the creator cuts. The landing-page scroll is FFmpeg zoompan over
43	a Playwright-rendered PNG — not an i2v clip (i2v hallucinates the UI).
44	- VO over a ducked bed. Mix the optional instrumental bed sidechain-ducked UNDER the VO
45	(−20dB, 20:1) so the VO stays clearly on top; the bed can drop in on the payoff beat.
46	- End card via the brand's real PNG — never AI-render brand text. Append the brand's real
47	end-card PNG (~2s) on the tail, captions suppressed. A diffusion model garbles a wordmark.
48	- FFmpeg composite, deterministic, FREE. Trim-to-EDL, `filter_complex concat`, VO+music mix,
49	caption burn, landing-page zoompan, end-card append, `loudnorm I=-14` → a 1080×1920 h264+aac
50	master (~37s). No paid c

People who liked this also viewed

6 picks

Type

render-narrated-ugc-wardrobe-stitch

1.0kstars193forks

View on GitHub

How to install render-narrated-ugc-wardrobe-stitch?

gooseworks-ai/goose-skills/render-narrated-ugc-wardrobe-stitch

$npx -y skills add gooseworks-ai/goose-skills --skill render-narrated-ugc-wardrobe-stitch

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# render-narrated-ugc-wardrobe-stitch
2
3	Assemble a narrated-UGC "stitch reply" ad from a config: a fast-cut vertical testimonial
4	where a single spoken VO carries a verbatim ~13-sentence reversal-hook monologue over ONE creator
5	across ~5 wardrobe changes in ~3 micro-worlds, interspersed with product B-roll (capsule macro,
6	unboxing, a landing-page scroll), ~30 hard cuts on the VO cadence, closing on a brand end card.
7	This capability is the FREE, deterministic assembly — trim-to-EDL, hard-concat, the VO+music
8	mix, the karaoke-pop caption burn, the landing-page zoompan, and the end-card append.
9
10	`scripts/config.example.json` is the worked example (Bioma "Do NOT buy Bioma Probiotics", ~37s
11	1080×1920 9:16, ~30 body cuts + a ~2s end card); `scripts/PIPELINE.md` maps every config block to
12	its source step and `scripts/README.md` documents the free assembly.
13
14	## Run
15
16	This is the FREE, deterministic assembly stage — it spends nothing. The paid inputs are
17	separate capabilities — the spoken VO (`create-vo-elevenlabs`) Whisper-aligned so the WORD
18	BOUNDARIES set the cut grid; one locked creator (`create-image-gpt-image-fal` anchor + ~5 wardrobe
19	edits chained off the anchor) + 3 world wides + per-cut start-frames (`create-image-fal` product
20	composites); and one Veo/Seedance i2v clip per cut (`create-video-fal`). Given the VO +
21	`vo-final.words.json` + `edl.json` + one clip per cut + a Playwright landing-page PNG + the brand
22	end-card PNG, `render-narrated-ugc-wardrobe-stitch` trims each clip to its EDL window, hard-concats
23	on the VO cadence, mixes the VO over the ducked bed, burns the karaoke-pop captions, appends the
24	end card → the master. Re-cuts reuse the existing VO / start-frames / clips and cost $0.
25
26	## Contract (the free assembly)
27
28	- The spoken VO carries the narrative — lock it FIRST. The VO IS the narration bed; the whole
29	ad is cut to it. Never plan the cut grid before the VO is locked and Whisper-aligned.
30	- Build the EDL from the VO's Whisper word boundaries. ~30 role-tagged cuts (`hook`, `feature`,
31	`reaction-insert`, `payoff-hold`, `b-roll-insert`, `landing-page`); snap every cut window to the
32	word boundaries. The payoff line gets a HELD `payoff-hold` beat (~3× mean shot length).
33	- Hard cuts via `filter_complex concat`, not the demuxer. Trim each clip to its EDL window and
34	hard-concat with `filter_complex concat` — the `-f concat` demuxer drops the audio when a
35	drawtext/scale step shaves a clip a few ms below its window. No dissolves.
36	- Karaoke-pop captions on every word, throughout. From the VO's `vo-final.words.json` (VEED
37	Whisper preset, bold yellow), on every word; re-spell brand tokens Whisper mishears against the
38	locked script ("synbiotic" over "symbiotic"; keep "I'ma" verbatim) — never edit the script to
39	match Whisper. Captions are suppressed over the end card. If VEED mis-captions a brand token,
40	hand-patch that sentence with local ASS karaoke.
41	- Product B-roll breaks up the talking head. Capsule macro, unboxing, and a landing-page
42	scroll are interspersed with the creator cuts. The landing-page scroll is FFmpeg zoompan over
43	a Playwright-rendered PNG — not an i2v clip (i2v hallucinates the UI).
44	- VO over a ducked bed. Mix the optional instrumental bed sidechain-ducked UNDER the VO
45	(−20dB, 20:1) so the VO stays clearly on top; the bed can drop in on the payoff beat.
46	- End card via the brand's real PNG — never AI-render brand text. Append the brand's real
47	end-card PNG (~2s) on the tail, captions suppressed. A diffusion model garbles a wordmark.
48	- FFmpeg composite, deterministic, FREE. Trim-to-EDL, `filter_complex concat`, VO+music mix,
49	caption burn, landing-page zoompan, end-card append, `loudnorm I=-14` → a 1080×1920 h264+aac
50	master (~37s). No paid c

People who liked this also viewed

6 picks

Type