Automate YouTube Videos with StoryForge using LangGraph and Azure OpenAI

This title was summarized by AI from the post below.

Been building something that I'm actually excited to share, instead of just thinking about my next job. I wanted to create moral story videos for YouTube — the kind with AI images, voiceover, and background music. Simple idea. But as a Python fullstack developer I thought... why manually do this when I can automate the whole thing? I already have experience in building Agentic AI architecture using langchain, Langgraph, RAG, so just brushing the skills. I built StoryForge. You paste a story. It gives you a YouTube-ready animated video. Here's what happens under the hood: → Azure OpenAI reads your story and splits it into scenes automatically (with image prompts for each scene) → LangGraph processes all scenes in parallel — images and voice generated simultaneously, not one by one → AI generates the visuals for free → edge-tts handles the narration → MoviePy stitches everything together — images synced to audio, subtitles, background music → FastAPI serves the whole thing with a clean UI and live progress streaming via SSE The part I enjoyed most was the LangGraph pipeline. Instead of processing 6 scenes sequentially (which would take forever), the Send API fans out all scenes in parallel. What used to take 3+ minutes now finishes in under 45 seconds. The entire pipeline costs roughly ₹2 per video (just the Azure OpenAI scene parsing call). Images and voice are free. Next up — adding emotional voice using Azure Speech Service so the narration actually feels like a story, not a robot reading text. Still a side project but it's been a good learning exercise, with no job in hand. Indeed the difficult time for every job seeker, but keeping yourself busy with something is the best thing to do. #Python #FastAPI #LangGraph #AzureOpenAI #GenerativeAI #SideProject #BuildInPublic #layoff

To view or add a comment, sign in

Explore content categories