Fallback to CPU FILM interpolation when GPU quota exceeded

Return 503 on GPU quota exceeded instead of 500

Fix gradio_client token param name

Use HF_TOKEN for authenticated ZeroGPU requests

Switch to LTX-2-3-First-Last-Frame to avoid ZeroGPU quota exhaustion

Fix /timelapse endpoint with correct Gradio API params

Add /timelapse endpoint using Wan 2.2 via gradio_client

Remove root GET endpoint

Redirect root to FastAPI docs

Remove index.html, return JSON from root endpoint

Add FastAPI app with Dockerfile for HF Spaces

Add index.html

initial commit