// the find
jakobhoeg/nextjs-ollama-llm-ui
Fully-featured web interface for Ollama LLMs
A Next.js chat front end for a local Ollama server, built with Tailwind, shadcn/ui and shadcn-chat. Conversations are stored in the browser's localStorage, so there is no database to run. It suits people who want a ChatGPT-style window over local models without running Open WebUI's heavier stack.
- No database and no accounts. Chats live in localStorage and the server side is thin, so deployment is one container with one required variable, OLLAMA_URL.
- It covers the Ollama features people use most from a UI: pulling and deleting models, switching models mid-conversation, image input for vision models, and regenerating a response.
- Code blocks get syntax highlighting and a copy button, which matters when most of the output is code.
- localStorage is the whole persistence story. Clearing site data, switching browsers, or hitting the browser's storage quota loses or breaks history, and the README still lists import and export as not done.
- The last push was June 2025 and the README calls this a hobby project, pointing people to Open WebUI for a fuller setup. Treat it as a clean reference implementation, not something with an active maintainer.
- No authentication. The Docker quick start publishes port 8080 on all interfaces, and anyone who can reach it can chat with your models and pull or delete them. Bind it to localhost or put a reverse proxy with auth in front.