Full Stack Web Development with AI
summary
Condensed demo builds a streaming chatbot with Next.js and the Vercel AI SDK, wiring a POST API route to stream tokens to a front-end useCompletion/useChat hook and get a working prototype in 2–3 minutes. Covers swapping providers by installing provider packages, explains model limits (Gemini 2.0 Oflash: 1,000,000 input tokens, 8,192 output), temperature/top_p and token math (~100 tokens ≈ 60–80 words), and enables live lookups via function calling. Ends with Cloudinary image work: generative fill, object removal, recolor and background removal.
AI-written summary · about
chapters (12)
- 0:00Introduction to AI talk
- 1:06What is a large language model
- 5:17Hallucinations in LLMs
- 7:22Next.js AI SDK overview
- 9:10Building a chatbot with Next.js
- 10:53Content generation types
- 11:56Model parameters explained
- 16:54Structured output with LLMs
- 17:58Live information with tool calling
- 20:24Switching between AI models
- 21:08AI image manipulation demo
- 23:10Q&A and resources
related talks
04mLightning Fast AI Chatbots in Nuxt
29mCoding Chat Bots
28mCreating fluent Frontend Experiences in an age of AI
57mSmart Chat App with Vue, AI and Serverless Function Solution LIVE
05mBuild a Smart Chat App with Vue, AI, and Serverless Function
21mUnlocking Client-Side AI in Your Web App
45mWeb Development in the Age of AI
01h 38mIntegrating AI into Your Apps: A Developer's Guide to Generative AI APIs