Repolode
Submit
HelgeSverre

HelgeSverre/ollama-gui

A browser-based chat interface for local LLMs served by Ollama, with chats stored in IndexedDB on your own machine.

Open-source alternative to:ChatGPT

View on GitHubWebsite
Stars
1,270
Forks
163
Open issues
2
Last commit
5 days ago

What is it?

Ollama GUI is a single-page web app that connects directly from the browser to an Ollama server. It has no backend of its own, so chats are kept in the browser's IndexedDB and only travel to the Ollama instance you configure.

The app supports streaming replies with stop, regenerate and edit actions, plus versioned edits and branching a reply into a new chat. It shows the reasoning of thinking models such as qwen3 and deepseek-r1 in a collapsible block, and accepts image attachments for vision models as well as text and code files.

Model management covers searching the Ollama library, pulling models with progress, inspecting capabilities and context length, and unloading models from memory. Generation parameters and system prompts can be set globally, per model or per chat, and chat history offers search, pinning, archiving and import/export as JSON or Markdown.

Key features

  • 01Streaming replies with stop, regenerate, and versioned message edits
  • 02Collapsible reasoning display for thinking models
  • 03Image and text/code attachments for vision and chat models
  • 04Search and pull models from the Ollama library with progress
  • 05Global, per-model, and per-chat generation settings
  • 06Chat history with full-text search, pinning, archiving, and JSON/Markdown export

Use cases

  • Chatting privately with models running on your own computer
  • Trying out different local models and comparing their replies
  • Using a local Ollama server on a home or office network from another device
  • Keeping a searchable, self-hosted archive of local LLM conversations

Quick start

  1. 1

    Pull a model with Ollama

    ollama pull llama3.2
  2. 2

    Start the Ollama server

    ollama serve
  3. 3

    Clone the repository

    git clone https://github.com/HelgeSverre/ollama-gui.git
  4. 4

    Install dependencies with Bun, then run bun run dev to open http://localhost:5173

    bun install

Things to know

  • Requires a running Ollama server, and development requires Bun and Node.js 20 or newer.
  • Service-worker offline caching and Gravatar avatars need HTTPS or localhost; plain HTTP LAN access has reduced features.

In collections

Self-hosted appsRun LLMs locallyPrivacy & local-first toolsDesktop GUI apps & toolkits

Topics

#ollama#llm#local-llm#chat-ui#self-hosted#privacy#vue#typescript#ai#gui

FAQ

What is ollama-gui?
A browser-based chat interface for local LLMs served by Ollama, with chats stored in IndexedDB on your own machine. Ollama GUI is a single-page web app that connects directly from the browser to an Ollama server. It has no backend of its own, so chats are kept in the browser's IndexedDB and only travel to the Ollama instance you configure.
Is ollama-gui free to use? What license does it have?
ollama-gui is open source under the MIT license. Check the license terms before using it in commercial products.
Is ollama-gui still maintained?
Actively maintained. The last commit to ollama-gui was on Nov 2, 2026.
What can I use ollama-gui for?
Chatting privately with models running on your own computer Trying out different local models and comparing their replies Using a local Ollama server on a home or office network from another device Keeping a searchable, self-hosted archive of local LLM conversations
How do I install ollama-gui?
Pull a model with Ollama: ollama pull llama3.2 Start the Ollama server: ollama serve Clone the repository: git clone https://github.com/HelgeSverre/ollama-gui.git Install dependencies with Bun, then run bun run dev to open http://localhost:5173: bun install
Where can I get ollama-gui?
The source code and installation instructions are on GitHub: https://github.com/HelgeSverre/ollama-gui

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects