ollama-chat-app

Ollama Chat App

A lightweight local web interface for chatting with Large Language Models (LLMs) that run entirely on your machine via Ollama. The default model is the multi-modal llava-llama3, but you can pull and switch to any other model exposed by Ollama.

✨ Features

• Real-time streaming chat with any Ollama model
• Image upload and paste support (works with multi-modal models such as llava-llama3)
• Built-in Model Manager – list, pull, delete and switch models from the UI
• Cancel in-flight requests
• Everything stays local – no data leaves your machine

📋 Prerequisites

1. Node.js ≥ 18
2. npm (comes with Node.js)
3. Ollama installed and available on your PATH (ollama serve should work)

🚀 Quick Start

The repository ships with a convenience script that spins everything up for you.

./start.sh

The script will:

1. Run ollama run llava-llama3 in the background (downloads the model on first run).
2. Wait for the Ollama REST API (http://localhost:11434) to become ready.
3. Launch the Node/Express server on port 3000.

Open http://localhost:3000 in your browser and start chatting!

Manual start

Already have Ollama running? Start the server directly:

ollama serve &   # if not already running
npm start        # installs dependencies if needed and starts the server

Set the environment variable PORT to override the default port (3000).

🗂️ Project Structure

ollama-chat-app/
├── public/           # Static front-end assets
│   ├── index.html    # Main HTML page
│   ├── app.js        # Chat UI logic
│   ├── models.js     # Model Manager modal
│   └── style.css     # Styles
├── server.js         # Express API ↔️ Ollama proxy
├── start.sh          # Convenience bootstrap script
└── package.json      # npm metadata & scripts

🔧 npm Scripts

• npm start – install dependencies (if missing) and start the app
• npm run dev – start with hot-reload via nodemon

📑 REST API Endpoints (served by server.js)

Method	Path	Description
GET	/api/models	List locally-available models
POST	/api/models/pull	Pull a model by name (streams progress)
GET	/api/models/:name	Fetch detailed model info
DELETE	/api/models/:name	Delete a local model
POST	/api/chat	Proxy to Ollama chat (streaming)

⚙️ Configuration

Variable	Default	Description
PORT	3000	Port for the Express server

If your Ollama instance is not running on localhost:11434, change the OLLAMA_API constant in server.js.

🛠️ Development

git clone https://github.com/isaaclins/ollama-chat-app.git
cd ollama-chat-app
npm install
npm run dev

The dev server uses nodemon for automatic reload on file changes.

📄 License

This project is licensed under the MIT License – see the LICENSE file for details.