[go: up one dir, main page]

Debian.Club

AI Tools

A comprehensive guide to installing and using AI tools on Debian — coding assistants, AI editors, local LLMs, image generation, and AI platforms.

AI tools have become an integral part of the modern development workflow — from terminal-based coding agents and intelligent IDEs to locally-run language models and image generators. This guide covers how to install and use the most popular AI tools on Debian.

AI Editors / IDEs

Code editors with built-in AI capabilities for intelligent completion, conversational programming, and code refactoring.

ToolDescriptionOpen SourceInstallation
CursorVS Code-based AI editor with built-in assistantNo.deb / AppImage
WindsurfAI Agent IDE by CodeiumNoAPT repository
ZedHigh-performance Rust editor with LLM integrationYesInstall script

AI Coding Assistants (CLI)

Terminal-based AI coding tools for command-line workflows.

ToolDescriptionOpen SourceInstallation
Claude CodeAnthropic's official terminal AI coding toolNoInstall script / npm
AiderAI pair programming CLI, multi-model supportYespip
OpenCodeOpen-source terminal AI coding agentYesInstall script / npm
ClineVS Code autonomous coding agent (5M+ users)YesVS Code extension
ContinueOpen-source AI code assistant (26K+ Stars)YesVS Code / JetBrains / CLI
GitHub CopilotGitHub's official AI coding assistantNogh CLI extension
Gemini CLIGoogle's open-source CLI AI tool, free for personal useYesnpm

AI Skills

Debian/Linux reliability rules that make AI agents verify system facts before installing packages, debugging services, or changing configuration.

EntryDescriptionStatus
DebianClub AI SkillsDebian/Linux environment detection, APT safety, command safety, systemd, networking, development setup, GPU, containers, security audits, and Debian packaging workflowsPhase 2
Install & DistributionLocal install, replace install, registry, packaging, and pre-release checksLive
UsageInvoke the skill in an AI agent and require fact verification before adviceLive
Safety BoundaryRead-only defaults, risky command detection, approval before changes, and log redactionLive

Local LLM Runners

Run large language models locally for privacy and offline use.

ToolDescriptionInstallation
OllamaMost popular local LLM runner with model registryInstall script
LM StudioGUI application for discovering and running LLMsAppImage
llama.cppHigh-performance C++ LLM inference engineBuild from source
llamafileSingle-file portable LLM, zero dependenciesDownload and run
JanPrivacy-first local AI chat applicationAppImage / .deb
LocalAIOpenAI API-compatible local inference serverDocker

AI Image & Speech

Locally-run AI image generation and speech recognition tools.

ToolDescriptionInstallation
Stable Diffusion WebUIMost popular AI image generation interfaceGit + Python
ComfyUINode-based AI image generation UIPython
OpenAI WhisperSpeech recognition and transcriptionpip

AI Platforms & Automation

Visual AI application builders and workflow automation tools.

ToolDescriptionInstallation
DifyLLM app development platform with visual workflow + RAGDocker
n8nWorkflow automation platform, 400+ app integrationsDocker / npm
OpenClawSelf-hosted AI agent platformInstall script

Common Prerequisites

Many AI tools share common dependencies. Install these before proceeding with individual guides.

Node.js

Claude Code, OpenCode, Gemini CLI, and similar tools usually require Node.js 18+. Check whether the Debian repository version satisfies the target tool first:

apt-cache policy nodejs npm

If the repository version is sufficient, prefer Debian packages:

sudo apt install nodejs npm
node --version
npm --version

If you truly need a third-party repository or installer script, review the source, GPG key, rollback path, and impact on the system Node.js package before making changes. Do not run unreviewed curl | bash commands.

Python

Required for Aider, Whisper, Stable Diffusion, and other ML tools.

# Debian 13 includes Python 3 by default
sudo apt install python3 python3-pip python3-venv

Docker

Recommended for Dify, n8n, LocalAI, and other platforms.

sudo apt install docker.io docker-compose-plugin
sudo usermod -aG docker $USER
# Log out and back in for changes to take effect

NVIDIA GPU Drivers

Optional but recommended for local LLMs and image generation.

# Install NVIDIA drivers
sudo apt install nvidia-driver

# Verify after reboot
nvidia-smi

Not all tools require a GPU. Ollama, llama.cpp, and others can run in CPU-only mode — just slower.

  1. Development Tools — Application installation on Debian
  2. Docker Guide — Docker installation and usage
  3. Command Line Basics — Getting comfortable with the terminal

Ready to boost your productivity? Start with Claude Code →

On this page