State isn't all you need, but It helps: building better LLM apps in the browser
summary
Demonstrates that small in-browser LLMs become practical when combined with concrete engineering: use WebLLM and Transformers.js for local models, apply few-shot prompting and RAG via a vector store to cut hallucinations, and orchestrate steps with LangGraph.js so workflows mix deterministic chains and limited agentic decisions. Shows a client-side prototype (LLAMA/vector store) and a Generative UI resume demo that runs in the browser, reduces formatting errors and hallucinations, and works reasonably well though not perfectly with current small models and Gemini Nano.
AI-written summary · about
chapters (4)
related talks
25mBuilding Your own AI Assistant with JavaScript, LLMs, and Rag
31mBringing and Running AI Agents in the Browser
34mMaking state management intelligent
33mDesigning LLM Applications for Massive Datasets
38m🔥 Generative UI: The Future of Frontend (And Why You Should Care)
12mNavigating the World of AI Development Without Going Insane
34mJavaScript-Powered AI: Crafting Intelligent Systems with LLMs and Vector Databases
09mIntegrating LLMs into Front-End Apps for Enhanced UX