LoveCode AI ⚡

Open-source terminal-native autonomous coding agent powered by free AI models.

LoveCode AI is a terminal-first autonomous coding assistant that helps developers write code, edit files, debug issues, execute commands, understand repositories, and automate workflows — using free and local AI models.

npm install -g lovecode-ai
lovecode

Features

| Feature | Description | |---------|-------------| | Lightweight | Fast startup and low RAM usage | | Open Source | Community-driven, MIT licensed | | Terminal Native | Keyboard-first workflow, no IDE required | | Free AI First | No paid API dependency | | Autonomous | Multi-step agent workflows | | Offline Capable | Local AI support via Ollama |

Install

npm (recommended)

npm install -g lovecode-ai

From source

git clone https://github.com/ashispluslaxmi-debug/LoveCode.git
cd lovecode
npm install
npm run build
npm link

Quick Start

# Start an interactive chat session
lovecode chat

# Run an autonomous task
lovecode run "refactor the auth module to use async/await"

# Initialize LoveCode in your project
lovecode init

Usage

lovecode chat

Start an interactive chat session with LoveCode AI.

lovecode chat                      # defaults to codellama via Ollama
lovecode chat -m deepseek-coder    # use a different model
lovecode chat -p ollama            # specify provider

lovecode run

Run LoveCode AI on a specific task in autonomous mode.

lovecode run "add input validation"
lovecode run "fix the failing tests" --dir ./packages/core

lovecode init

Initialize LoveCode AI configuration in your project.

lovecode init
lovecode init --force    # overwrite existing config

AI Providers

| Provider | Free? | Offline? | Default Model | |----------|-------|----------|---------------| | Ollama | ✅ | ✅ | codellama | | OpenAI Compatible | ✅ | ❌ | gpt-4o-mini |

LoveCode prioritizes free and local providers by default.

Prerequisites for local models

1. Install Ollama
2. Pull a model: ollama pull codellama
3. Run LoveCode: lovecode chat

Configuration

LoveCode looks for config in .lovecode/config in your project or ~/.config/lovecode/.

{
  "provider": "ollama",
  "model": "codellama",
  "temperature": 0.2,
  "maxTokens": 4096
}

Architecture

lovecode/
├── bin/              # CLI entry point
├── src/
│   ├── commands/     # Command implementations
│   ├── core/         # Agent and tool system
│   ├── ai/           # AI provider integrations
│   └── utils/        # Shared utilities
├── dist/             # Compiled output
└── .lovecode/        # Project config

Roadmap

• [x] CLI framework with chat, run, and init commands
• [ ] File read/write/edit tools
• [ ] Shell command execution
• [ ] Multi-step autonomous agent loop
• [ ] Code search (grep, glob)
• [ ] Git integration
• [ ] Multiple provider support
• [ ] Plugin system
• [ ] MCP server support
• [ ] Windows/WSL/Termux support

Requirements

• Node.js 18+
• Ollama (for local AI) or an OpenAI-compatible API

Development

git clone https://github.com/anomalyco/lovecode.git
cd lovecode
npm install
npm run dev           # watch mode
npm run build         # production build
npm run test          # run tests
npm run lint          # lint code
npm run typecheck     # type check

Why LoveCode?

Most AI coding assistants are:

• Paid — requiring expensive subscriptions
• Heavy — running inside resource-hungry IDEs
• Closed — proprietary, with no community input
• Tied to the cloud — requiring internet always-on

LoveCode AI is none of those. It runs in your terminal, uses free/local AI, and is fully open source.

License

MIT — see LICENSE