README.md

metadata

title: Product Showcase Studio
emoji: 🎬
colorFrom: gray
colorTo: indigo
sdk: docker
app_port: 7860

Product Showcase Studio

Cinematic product video pipeline: shot planning → KIE Veo / Seedance → FFmpeg merge, with optional GPT Image keyframes and product URL import.

Deploy on Hugging Face Spaces

1. Create a Space → Docker → link this repository (or push these files to a new repo under your account).
2. Settings → Repository secrets (or Variables and secrets): add at least:
   • VITE_API_BASE_URL — your Space’s public app URL, same as the address in the browser when the app loads, with no trailing slash (for example https://YOUR-USERNAME-YOUR-SPACE-NAME.hf.space). This is required so callback URLs and hosted images use the correct origin.
   • KIE_API_KEY — if you use KIE Veo.
   • OPENAI_API_KEY — optional, for LLM shot plans and GPT Image keyframes.
   • REPLICATE_API_TOKEN — optional Seedance fallback when KIE is not used.
3. Wait for the build (Node build + Python image). Open the Space URL when the build is Running.

The container runs uvicorn on PORT (Spaces set this; default 7860). FFmpeg is installed for the merge step.

Local Docker (optional)

docker build -t product-showcase .
docker run --rm -p 7860:7860 \
  -e VITE_API_BASE_URL=http://127.0.0.1:7860 \
  -e KIE_API_KEY=... \
  product-showcase

Then open http://127.0.0.1:7860.

Local development

See FLOW.md for architecture and env details.

# API
cd backend && python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt && python main.py

# UI (separate terminal)
cd frontend && npm install && npm run dev

With a built frontend/dist, python main.py also serves the SPA from the same origin.