omni-video-director

skill
Guvenlik Denetimi
Gecti
Health Gecti
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 14 GitHub stars
Code Gecti
  • Code scan — Scanned 7 files during light audit, no dangerous patterns found
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Drop in a video — Claude adds the b-roll & VFX for you. Google Omni + GPT-Image-2 via kie, assembled with HyperFrames.

README.md

🎬 Omni Video Director

Drop in a video and AI adds the graphics and effects for you — turn a bottle transparent, make a little mascot climb on your shoulder, drop in animated stat cards, give yourself a beard. You pick where; the AI does the rest. You bring the video, it brings the effects.

What you'll need

Three sign-ups + a couple of free tools. All spelled out below — it's fine if these are new to you.

  • Claude Code — the app you'll "talk" to (it runs everything).
  • A kie.ai account — lets Claude use Google Omni to add graphic effects to your work. Pay-as-you-go: load a few dollars; each effect costs a little.
  • An OpenRouter account — lets you access models like Google Gemini 3.5 Flash to analyse your videos. Costs pennies.
  • ffmpeg + jq — free background tools that stitch and read your files. Install once. (Node is optional — only if you later switch on the kie MCP; see QUICKSTART.)

Part 1 — One-time setup (~10 min, only done once)

1. Install Claude Codehttps://claude.com/claude-code (follow their installer).

2. Install the background tools. Open the Terminal app (Mac: press ⌘+Space, type "Terminal", Enter), paste this, press Enter:

brew install ffmpeg jq

(No "brew"? Install it first from https://brew.sh. On Windows: get ffmpeg from ffmpeg.org and jq from jqlang.github.io/jq.)

3. Get the kit. On this GitHub page click the green "Use this template" → Create a new repository (or just Code → Download ZIP and unzip it).

4. Get your 2 keys. A "key" is a password that lets the kit use a service (and bills you for what you use).

  • kie: kie.ai → sign up → Dashboard → API Keys → copy. Then add a few dollars of credit.
  • OpenRouter: https://openrouter.ai/keys → sign up → Create Key → copy.

5. Paste your keys in. In the kit folder, copy the file .env.example and name the copy .env. Open it and paste your keys after the =:

KIE_API_KEY=your-kie-key-here
OPENROUTER_API_KEY=your-openrouter-key-here

Save it. Setup done — you never do this part again.
(Shortcut: open the folder in Claude Code and say "help me set up my .env" — it'll walk you through it.)

Part 2 — Make a video (every time)

6. Film a short clip on your phone — you talking, holding a prop, gesturing. It transforms real footage, so give it something real. (See projects/_TEMPLATE/input/WHAT-TO-FILM.md for ideas.)

7. Start a project. Open the folder in Claude Code and say:

"Start a new project called my-first-video"

It makes a folder and tells you where to drop your clip. Put your video there.

8. Let it direct. Say:

"Act as my video director."

Then it's a simple back-and-forth:

  1. It watches your video and suggests where effects go → you approve or edit the list.
  2. It writes the exact instructions for each effect.
  3. It shows the cost → you say "go" → it makes the effect.
  4. You watch each one → keep what you like.
  5. It stitches your finished video into the output/ folder.

Good to know

  • It never spends your money without asking — you approve every effect first.
  • Costs are small — usually a fraction of a dollar per effect, shown before each one.
  • You're the director — if something's off, just say so and it redoes it.

Advanced / optional (the cleaner "kie MCP" setup, long-video analysis, troubleshooting): see QUICKSTART.md.

Built by Dainami AI. Want the whole system that runs a business on AI agents? → the AI OS Blueprint.

Yorumlar (0)

Sonuc bulunamadi