In this video I build a PDF summarizer using AI and large language models, and I do it two ways: first as a super simple Node.js CLI, then as a clean Next.js + React UI. The CLI version is perfect if you want something quick that runs in the terminal: I drop PDFs into a /files folder, use prompts (with ys) to pick the PDF and the OctoAI model, parse the PDF into plain text, send it to the model, then export the summary back into the same folder as a .txt file. After that, I take the exact same idea and turn it into an actual web app with Next.js. On the client I use pdf.js to extract the text (so the heavy lifting isn’t all on the server), then I POST it to an API route that calls OctoAI for summarization and returns the summary to display in the UI. I also talk through the “real world” stuff: basic error handling, keeping your API key in env, and how to chunk large PDFs into smaller pieces so you can summarize big documents and then combine those chunk summaries into one “super summary.”