Getting Started

Getting Started

Prerequisites

• Node.js 22+ — Download
• pnpm 10+ — npm install -g pnpm
• A local LLM — Ollama, LM Studio, llama.cpp, or vLLM

Quick Start (Browser)

git clone https://github.com/wildfirebill-ai/localai-code-editor.git
cd localai-code-editor
pnpm install
pnpm build
pnpm dev:server   # Terminal 1 — starts the backend
pnpm dev:web      # Terminal 2 — starts the frontend

Open http://localhost:5173 in your browser.

Quick Start (Electron Desktop)

pnpm dev:desktop  # Starts Electron with hot reload

Quick Start (Docker)

docker compose up

Open http://localhost:4801.

First Run

1. Set a workspace — click "Open Folder" or type a path in Settings
2. Add an LLM provider — Ollama preset is pre-configured for localhost:11434
3. Start chatting — click the ✨ button and ask the AI about your code

LLM Provider Setup

Ollama (Recommended)

# Install and start Ollama
curl -fsSL https://ollama.com/install.sh | sh
ollama serve

# Pull a model
ollama pull codellama:7b

The editor auto-detects Ollama at http://localhost:11434/v1.

LM Studio

1. Download LM Studio
2. Load a model and start the local server
3. The editor connects at http://localhost:1234/v1

llama.cpp

# Build and run
cmake -B build
cmake --build build --config Release
./build/bin/llama-server -m model.gguf

The editor connects at http://localhost:8080/v1.

What's Next?

• Configuration — customize providers, MCP servers, and editor settings
• Skills — enable agent skills for automated workflows
• [MCP Servers](MCP Servers) — connect external tools