express-ollama-server

A Node.js Express server with a local LLM chat endpoint powered by Ollama.

Endpoints

Method	Path	Description
GET	/health	Health check
GET	/users	Returns a list of users
POST	/chat	Send a prompt to the local LLM

POST /chat

Request

{ "prompt": "Write a function to reverse a string" }

Response

{
  "model": "qwen2.5-coder:7b",
  "response": "..."
}

Setup

Prerequisites: Node.js 18+, Ollama installed and running.

# Install dependencies
npm install

# Pull the model
ollama pull qwen2.5-coder:7b

# Start the server
npm start

Server runs on http://localhost:3000 by default.

Configuration

Env var	Default	Description
PORT	3000	Server port
OLLAMA_URL	http://localhost:11434	Ollama base URL
OLLAMA_MODEL	qwen2.5-coder:7b	Model to use

Testing

npm test

Tests use Jest + supertest. The Ollama API is mocked so tests run without a live model.