NeuroByte

AI-Powered Local Coding Assistant which runs entirely on your machine using Ollama.

NeuroByte is a VS Code extension that provides Claude Code-like capabilities using local LLM models through Ollama. It offers full project access, plan mode, automatic file editing, and intelligent code assistance - all running privately on your device.

Features

• Chat Interface - Conversational AI assistant integrated into VS Code
• Plan Mode - Create and execute multi-step implementation plans
• Automatic File Editing - AI can read, write, and edit files directly
• Project Indexing - Fast semantic search across your codebase
• Context-Aware - Understands your current file, selection, and project structure
• Device-Optimized - Automatically recommends models based on your hardware
• Ollama Integration - Auto-install, auto-start, and auto-pull models
• Fully Local - No data leaves your machine

Quick Install

macOS / Linux

curl -fsSL https://raw.githubusercontent.com/spyxpo/neurobyte/stable/scripts/install.sh | bash

Windows (PowerShell)

iwr -useb https://raw.githubusercontent.com/spyxpo/neurobyte/stable/scripts/install.ps1 | iex

npm

npm install -g @spyxpo/neurobyte

Homebrew (macOS/Linux)

brew tap neurobyte/tap
brew install neurobyte

VS Code Extension

Search for "NeuroByte" in the VS Code Extensions marketplace, or:

code --install-extension neurobyte.neurobyte

Requirements

• Ollama - Local LLM runtime (auto-installed by the installer)
• VS Code 1.85.0 or later
• Recommended: 8GB+ RAM (works with less using smaller models)

Model Recommendations

NeuroByte automatically detects your device capabilities and recommends appropriate models:

Device Tier	RAM	Recommended Model	Alternatives
Low	< 8GB	qwen2.5:0.5b	tinyllama, phi3:mini
Medium	8-16GB	qwen2.5:3b	phi3, llama3.2:3b, codellama:7b
High	16-32GB	qwen2.5:7b	llama3.1:8b, codellama:13b
Ultra	32GB+	qwen2.5:14b	llama3.1:70b, codellama:34b

Usage

Chat

1. Open the NeuroByte panel from the sidebar (brain icon)
2. Type your question or request
3. NeuroByte will respond and can execute actions on your codebase

Quick Actions

• Cmd/Ctrl + Shift + N - Open NeuroByte chat
• Cmd/Ctrl + Shift + A - Ask about selected code
• Right-click context menu for code actions

Plan Mode

1. Run "NeuroByte: Enter Plan Mode" from command palette
2. Describe what you want to accomplish
3. Review the generated plan
4. Approve to execute step by step

Commands

Command	Description
NeuroByte: Open Chat	Open the chat panel
NeuroByte: Ask About Selection	Ask about selected code
NeuroByte: Explain This Code	Get explanation of selected code
NeuroByte: Fix This Code	Find and fix issues
NeuroByte: Refactor This Code	Improve code quality
NeuroByte: Generate Tests	Create unit tests
NeuroByte: Enter Plan Mode	Start planning mode
NeuroByte: Select Model	Choose a different model
NeuroByte: Pull Model	Download a new model
NeuroByte: Index Project	Index current workspace
NeuroByte: Setup Ollama	Install/start Ollama

Configuration

Settings can be configured in VS Code settings (Cmd/Ctrl + ,):

{
  "neurobyte.ollamaUrl": "http://127.0.0.1:11434",
  "neurobyte.defaultModel": "qwen2.5:7b",
  "neurobyte.autoStartOllama": true,
  "neurobyte.autoPullModel": true,
  "neurobyte.streaming": true,
  "neurobyte.temperature": 0.7,
  "neurobyte.maxContextLength": 8192,
  "neurobyte.planMode.requireApproval": true,
  "neurobyte.excludePatterns": [
    "**/node_modules/**",
    "**/.git/**"
  ]
}

Architecture

NeuroByte consists of two components:

1. NeuroByte Core (Rust) - Fast backend for file operations, indexing, and Ollama communication
2. VS Code Extension (TypeScript) - UI and editor integration

┌─────────────────────────────────────────────────────┐
│                    VS Code                          │
│  ┌─────────────────────────────────────────────┐    │
│  │           NeuroByte Extension               │    │
│  │  ┌─────────┐ ┌─────────┐ ┌─────────────┐   │    │
│  │  │  Chat   │ │  Plan   │ │   Models    │   │    │
│  │  │  View   │ │  View   │ │    View     │   │    │
│  │  └────┬────┘ └────┬────┘ └──────┬──────┘   │    │
│  └───────┼───────────┼─────────────┼──────────┘    │
│          └───────────┼─────────────┘               │
│                      ▼                              │
│  ┌─────────────────────────────────────────────┐    │
│  │           NeuroByte Core (Rust)              │    │
│  │  ┌─────────┐ ┌─────────┐ ┌─────────────┐   │    │
│  │  │  File   │ │ Project │ │    Plan     │   │    │
│  │  │  Ops    │ │ Indexer │ │   Engine    │   │    │
│  │  └─────────┘ └─────────┘ └─────────────┘   │    │
│  └───────────────────┬──────────────────────────┘   │
│                      ▼                              │
│  ┌─────────────────────────────────────────────┐    │
│  │                 Ollama                       │    │
│  │           (Local LLM Runtime)               │    │
│  └─────────────────────────────────────────────┘    │
└─────────────────────────────────────────────────────┘

Building from Source

Prerequisites

• Rust 1.75+
• Node.js 18+
• npm or yarn

Build

# Clone the repository
git clone https://github.com/spyxpo/neurobyte.git
cd neurobyte

# Build everything
./scripts/build.sh

# Or build components separately
./scripts/build.sh --core      # Build Rust backend
./scripts/build.sh --extension # Build VS Code extension
./scripts/build.sh --npm       # Create npm package
./scripts/build.sh --brew      # Create Homebrew formula

Development

# Start the core server
cd neurobyte-core
cargo run

# In another terminal, watch the extension
cd neurobyte-extension
npm install
npm run watch

# Press F5 in VS Code to launch Extension Development Host

Author

• Mantresh Khurana - Creator and maintainer