Fix app.py: use DEFAULT_MODEL (gemma3:12b) for dropdown, free-tier info label

Fix 403: free-tier models first in dropdown, default=gemma3:12b, friendly error messages for subscription/auth/rate-limit

Update app.py: cloud models dropdown, custom model input, optional image API keys, multi-source images

Update media_downloader: multi-source images (Pexels + Openverse + Pixabay + Picsum)

Update: Real Ollama cloud models from ollama.com/search, custom model input, multi-source images, optional API keys

Fix: robust FFmpeg renderer with drawtext auto-detection and proper text escaping

Fix: add lxml_html_clean dep for trafilatura + robust FFmpeg renderer with drawtext detection

Fix Dockerfile: handle existing UID 1000 from node base image

Add app.py - Main Gradio application with complete pipeline

Add video_render.py - Remotion + FFmpeg video rendering

Add media_downloader.py - Pexels stock image/video downloader

Add tts.py - edge-tts voice generation with pydub

Add llm.py - Ollama cloud script generation

Add Python modules: scraper, llm, tts, media_downloader, video_render

Add Remotion ViralVideo.tsx with transitions and effects

Add Remotion Root.tsx

Add Remotion source: index.ts

Add public/.gitkeep

Add remotion.config.ts

Add tsconfig.json and remotion.config.ts

Add package.json

Add requirements.txt

Add Dockerfile

Add README

initial commit