Build local and offline-capable chatbots
Large language models (LLMs) open the door to chatbots that generate content, summarize text, analyze sentiment, and hold natural language conversations. Newer applications like Gemini and ChatGPT rely on advanced natural language processing (NLP), but running that logic entirely in the cloud is not the only path. Google Developer Expert Christian Liebel has written a series for web.dev on adding AI-powered chat directly to websites and web apps so they can work locally and even offline.
The series covers three practical steps:
- Why client-side LLMs are worth the switch for user privacy, cost, and latency.
- How to add a WebLLM-powered chatbot to a to-do list application, with AI executing entirely in the browser.
- How to use the built-in Prompt API to experiment with browser-native AI without external dependencies.
If you'd like to share what you build or request office hours with the Web.dev AI team, reach out to @ChromiumDev on X.
