Skip to content

gooseworks-ai/goose-skills / skills/video/render-brand-identity-reveal

render-brand-identity-reveal

Verified

Render a 'brand identity reveal' video from a config — a single poster frame in a real, softly-lit space (real wall, soft-focus plant in the corner, dappled leaf shadow, illuminated poster) whose artwork HARD-CUTS through ~10 on-brand poster mockups (hero product, IG post, hanging banners, sticker sheet, logo lockup, poster, two lifestyle stills, packaging, big icon) then holds on a brand end card. The environment plate is one create-image-fal generation; the mockups are real-DOM HTML frame-stepped via Playwright, perspective-composited into the detected frame quad with the plate's real leaf-shadow multiplied back onto each poster (reads as behind glass), sequenced by FFmpeg. Deterministic assembly, FREE (the plate comes from create-image-fal, the bed from create-music-elevenlabs), music bed only and approved brand copy only. Use for the brand-identity-reveal format.

FileKindSize
SKILL.mdskill md5.3 KB
scripts/PIPELINE.mdmarkdown3.5 KB
scripts/build_video.shexecutablescript1.2 KB
scripts/composite.pyscript3.0 KB
scripts/config.example.jsonconfig3.9 KB
scripts/measure_frame.pyscript1.7 KB
scripts/recrop.pyscript2.2 KB
scripts/render_art.pyscript1.3 KB
scripts/scene.htmlother20.4 KB
skill.meta.jsonconfig223 B
tests/smoke-test.mdmarkdown1.1 KB

scripts/PIPELINE.md

View on GitHub
# PIPELINE — brand-identity-reveal free assembly

The free renderer is **documentation-grade**: `scene.html` is brand-specific (the shipped
reference is the Palmglow instance), and the Python scripts drive Playwright + PIL + FFmpeg
around it. Adapt `scene.html` per brand; the scripts are generic.

## Working-dir layout (per project)

```
working/
  scene.html            # 11 poster modules; one active via `.on` class (toggle by OPACITY)
  assets/products/*     # the brand's real product / lifestyle stills
  assets/brand/*        # wordmark png, brand-icon svg
  bg/plate_1.png        # PAID: create-image-fal environment plate (high-res, empty frame)
  bg/plate_final.png    # recrop.py output (camera distance)
  bg/corners.txt        # measure_frame.py output (the frame interior quad)
  art/art_std_01..11.png# render_art.py output (standalone posters, bare mode, 2x)
  frames_v2/frame_01..11.png  # composite.py output (posters warped into the lit frame)
  concat.txt            # ffmpeg concat list (frames + per-beat durations)
  music.mp3             # PAID: create-music-elevenlabs bed
```

## Steps (script → output)

1. **Plate [PAID]** — `create-image-fal` (flux-pro ultra, 9:16) → `bg/plate_1.png`.
   `gen_plate.py` is the reference wrapper (loads FAL_KEY, prompt in `config.plate`; `--dry-run`
   prints the prompt). `plate.setting` and `plate.wall_style` come from the recipe's `setting` and
   `wall_style` choices (the Palmglow demo used a boutique interior + sage subway tile — an
   example, not a default). Tune the wall colour to complement the brand's dominant hue so the
   posters pop.
2. **Camera** — `recrop.py <interior_width_frac> <top_frac> [plate.png]` → `bg/plate_final.png`
   (source defaults to the first `bg/plate_*.png`). The plate
   is high-res, so cropping closer costs no quality. `0.72 0.13` = frame ~72% of width.
3. **Detect frame** — `measure_frame.py` → `bg/corners.txt` (bright + low-sat + near-neutral
   mask → largest component → 4 corners).
4. **Mockups** — author `scene.html` (11 modules from the brand's assets + approved copy), then
   `render_art.py` → `art/art_std_01..11.png` (Playwright bare mode: `body.bare` hides the CSS
   room and the frame fills the viewport; `device_scale_factor=2`).
5. **Composite** — `composite.py`: perspective-warp each poster into the quad, **multiply the
   plate's interior luma back onto it** (`shadow_strength`), glass sheen → `frames_v2/*`.
6. **Sequence** — write `concat.txt` (frames + `duration` lines per `config.beats[].seconds`,
   repeat the last frame line), then `build_video.sh concat.txt music.mp3 out.mp4 <total_s>`.

## Adapting `scene.html` per brand

- Keep the 11 module ids (`.a1`..`.a11`, `data-i`), the `#thand` icon symbol pattern, and the
  bare-mode CSS. Swap: product image paths, wordmark/icon, palette CSS vars, and the baked
  copy (approved only). Recreate the brand icon as an SVG `<mask>` if the only source has an
  occluding element.
- Beats are opaque full-bleed; **toggle with opacity, not display** (a per-state `display:flex`
  silently overrides a `display:none` toggle).

## Split (per adding-and-testing-a-video-format.md)

- **Paid** (already generic caps): `create-image-fal` (plate), `create-music-elevenlabs` (bed).
- **Free renderer** (`render-brand-identity-reveal`): `scene.html`, `render_art.py`,
  `recrop.py`, `measure_frame.py`, `composite.py`, `build_video.sh`. (`gen_plate.py` is a
  convenience wrapper around the paid `create-image-fal` step — not part of the free cap.)
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 323kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 323kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 323kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall