fmaclen/hollama

What it solves

Hollama is a minimal chat interface that allows users to interact with Large Language Models (LLMs) without needing a complex setup or a dedicated backend server for the UI. It runs entirely in the browser, meaning the application logic and data storage are handled locally on the user's machine.

How it works

The app connects to LLM servers such as Ollama and OpenAI. It supports multi-server configurations, allowing users to switch between different providers. It handles both text and vision models, as well as reasoning models, and renders the output using Markdown, KaTeX for math, and syntax highlighting for code.

Who it’s for

Users who want a lightweight, privacy-focused chat interface for their local or remote LLM servers (like Ollama) or API-based models (OpenAI) without requiring a sign-up process.

Highlights

• own-browser execution: The app runs entirely in the browser, and data is stored locally.
• Multi-server support: Compatible with both Ollama and OpenAI servers.
• Rich rendering: Supports Markdown, KaTeX math notation, and code editor features.
• Model management: Allows users to download Ollama models directly from the interface.
• Session management: Includes features to edit and retry messages, and the ability to import and export stored data.