← ClaudeAtlas

satva-ledger-marketing-videolisted

Reference pipeline to FORK for a longer "sizzle" / hero marketing video — the Satva Ledger (AI accounting agent) 82-second promo. A generator script builds the HyperFrames composition from data, an original SVG mascot animates it, and a Python audio pipeline makes the ElevenLabs voice + SFX, an original synthesised score, the ducked mix and measured QA. Use when asked to rebuild, re-render or adapt the Satva Ledger / accounting-agent marketing video, or to make a multi-chapter promo with a mascot and beat-synced sound design. For a quick launch video of ANY single feature, use feature-launch-video instead.
prafullakalu/Satva-COE-skills-agents · ★ 0 · Data & Documents · score 67
Install: claude install-skill prafullakalu/Satva-COE-skills-agents
# Satva Ledger marketing video — the reference pipeline This is the **real, hand-built pipeline** behind the Satva Ledger promo, kept readable so the team can rerun it or fork it for the next product. It is **not** a portable template: the chapters, the 25 voice lines and the 82.567 s length are written into the scripts on purpose. ## Which video skill do I use? | | `feature-launch-video` | **`satva-ledger-marketing-video`** (this) | |---|---|---| | Length / shape | 25–45 s, one feature | ~80 s, intro → 6 chapters → all-in-one → end card | | Input | a `brief.json` — no code edits | you edit the generator and the voice script | | Sound | generic, timing-driven | hand-cued so key words land on visual beats | | Portable | yes — any product | no — a worked reference to fork | If you only need "a launch video for feature X", stop and use `feature-launch-video`. ## What is in the folder ``` pipeline/ video-build/ build-sizzle.mjs generates satva-sizzle/index.html from data (chapters, timings, cues) sizzle-mascot.mjs the original SVG mascot + mood/bob/blink/wave/hop helpers (brand palette only) satva-sizzle/ composition scaffold: fonts (OFL), official Satva logo, hyperframes config audio-pipeline/ el.py ElevenLabs helper (key from env or .env, never logged) + ffmpeg decode gen_voice.py 25 voice lines, silence-trimmed, placed by startAt/endAt on the visual beats gen_sfx.py 10 custom sou