VS Code Extension

AI Coding
Assistant.
Local. Free.

NEURO brings AI-powered code explanation, bug fixing, test generation, and refactoring to VS Code — powered by Ollama, running 100% on your machine.

Install Extension Watch Demo
utils.ts — neuro-vscode
utils.ts
app.ts
1
function calculateTotal(items: CartItem[]) {
2
let total = 0;
3
for (const item of items) {
4
total += item.price * item.quantity;
5
}
6
return total;
7
}
8
9
Ln 4, Col 38 TypeScript
NEURO: Ollama connected mistral
0+
Downloads
0%
Local & Private
€0
API Costs Forever
0+
AI Models Supported

The Problem

Cloud AI assistants are expensive (€10-20/month), introduce latency, and require sending your proprietary code to third-party servers. Your code, your secrets, your logic — all leaving your machine.

The Solution

NEURO runs entirely on your machine via Ollama. Zero API costs, zero data leaks, zero latency to external servers. Your code never leaves your device.

Powerful Features

Everything you need, built in

Explain Code

Select any code and get a clear, detailed explanation of what it does.

Fix Code

Detect and fix bugs automatically with AI-powered analysis.

Generate Tests

Auto-generate comprehensive test suites for your functions.

Refactor

Improve code structure, naming, and patterns intelligently.

AI Chat

Chat with AI about your codebase directly in VS Code.

View all features

Built With Leading Technology

Multi-provider AI infrastructure

VS Code VS Code
TypeScript TypeScript
OpenAI OpenAI
Anthropic Claude AI
Node.js Node.js
GitHub GitHub

Comparison

How NEURO stacks up

NEURO AI GitHub Copilot Cursor Cody
Price Free forever €10/mo €20/mo Free tier + €9/mo
Privacy 100% local Cloud (Microsoft) Cloud Cloud (Sourcegraph)
API Costs €0 Included Usage-based Rate limited
Works Offline Yes No No No
Code Explanation Yes Yes Yes Yes
Bug Fixing Yes Limited Yes Yes
Test Generation Yes Yes Yes Yes
Custom Models Any Ollama model GPT-4 only Multiple Claude only
Open Source Yes (MIT) No No Partial

See It In Action

AI inline completions, right in your editor

Watch NEURO suggest code as you type. Accept with Tab, dismiss with Escape.

server.py — neuro-vscode
server.py
1
from flask import Flask, jsonify, request
2
3
app = Flask(__name__)
4
5
@app.route('/api/users', methods=['GET'])
6
def get_users():
7
8
9
Ln 7, Col 5 Python
NEURO mistral
See full interactive demo

Get Started

Up and running in 3 steps

1

Install Ollama

Download and install Ollama, then pull a model:

Terminal
$ curl -fsSL https://ollama.ai/install.sh | sh
$ ollama pull mistral
pulling manifest... done
2

Download the extension

Install the VSIX file in VS Code:

Terminal
$ code --install-extension neuro-ai-0.1.0.vsix
Extension 'neuro-ai' v0.1.0 was successfully installed.

Or: Download neuro-ai-0.1.0.vsix and install via Extensions > Install from VSIX

3

Start coding with AI

Open the command palette and use any NEURO command:

> NEURO:
NEURO: Explain Selection
NEURO: Fix Selection
NEURO: Generate Tests
NEURO: Refactor Selection
NEURO: Open Chat Cmd+Shift+N

FAQ

Frequently asked questions

What do I need to run NEURO?
You need VS Code and Ollama installed on your machine. Ollama runs AI models locally and is available for macOS, Linux, and Windows. NEURO connects to Ollama's local API to provide AI features. Any machine with 8GB+ RAM can run smaller models like Mistral.
Is my code sent to any external server?
No. NEURO communicates exclusively with the local Ollama server running on your machine (localhost:11434). No code, prompts, or data ever leave your device. There is no telemetry, no analytics, no cloud component whatsoever.
Which AI models can I use?
Any model available through Ollama. Popular choices include Mistral (fast, general purpose), CodeLlama (optimized for code), Llama 3 (latest Meta model), and DeepSeek Coder. You can switch models at any time in VS Code settings.
Will NEURO always be free?
Yes. NEURO is MIT licensed and will remain free and open source. The optional sponsorship tiers are purely for those who want to support development — they do not unlock any additional features.
How does NEURO compare to GitHub Copilot?
GitHub Copilot costs €10/month, requires an internet connection, and sends your code to Microsoft's servers. NEURO is free, works offline, and keeps everything local. While Copilot may have an edge in raw model quality (GPT-4), NEURO gives you full control over your models and complete privacy.