Local Mellum - Local AI Code Completion for VS Code

Local Mellum is a Visual Studio Code extension that provides AI-powered code completion using a local language model through Ollama. It helps you write code without sending your data to external servers.

Features

• Local AI-powered code completion
• Context-aware suggestions based on your codebase
• Privacy-focused - all processing happens on your machine
• No internet connection required for operation

Requirements

• Running Ollama with the proper model downloaded (see Setup Guide below)
• Your machine should follow the default requirements for local Mellum execution: https://www.jetbrains.com/help/ide-services/jetbrains-mellum.html#mellum-requirements

Setup Guide by Platform

Linux

# Install Ollama
curl -fsSL https://ollama.ai/install.sh | sh

# Start Ollama service
ollama serve

# Pull the required model
ollama pull JetBrains/Mellum-4b-sft-all

# Verify Ollama is running
curl http://localhost:11434/api/tags

macOS

# Install Ollama using Homebrew
brew install ollama

# Start Ollama service
ollama serve &

# Alternatively, download and install from the official website:
# https://ollama.ai/download/mac
# Then start from Applications folder

# Pull the required model
ollama pull JetBrains/Mellum-4b-sft-all

# Verify Ollama is running
curl http://localhost:11434/api/tags

Windows

1. Install from the official website: https://ollama.ai/download/windows
2. Ollama service should start automatically
3. Pull the required model (through CMD)

ollama pull JetBrains/Mellum-4b-sft-all

Configuration

Model Selection

By default, Local Mellum uses the JetBrains/Mellum-4b-sft-all model. You can configure a different model in your VS Code settings:

1. Open VS Code Settings (File > Preferences > Settings or Ctrl+,)
2. Search for "Local Mellum"
3. Set the "Model Name" to the name of your preferred Ollama model

{
  "localMellum.modelName": "JetBrains/Mellum-4b-sft-all"
}

4. Execute local-mellum.restart command to apply changes.

Make sure the model you specify is available in your Ollama installation. You can pull a new model using:

ollama pull your-model-name

License

All third-party licenses are available in LICENSE.md

Feedback and Support

TODO