tlm - Local CLI Copilot, powered by Ollama. 💻🦙

tlm is your CLI companion which requires nothing except your workstation. It uses most efficient and powerful open-source models like Llama 3.3, Phi4, DeepSeek-R1, Qwen of your choice in your local environment to provide you the best possible command line assistance.

Get a suggestion	Explain a command

Ask with context (One-liner RAG)	Configure your favorite model

Features

• 💸 No API Key (Subscription) is required. (ChatGPT, Claude, Github Copilot, Azure OpenAI, etc.)

• 📡 No internet connection is required.

• 💻 Works on macOS, Linux and Windows.

• 👩🏻‍💻 Automatic shell detection. (Powershell, Bash, Zsh)

• 🚀 One liner generation and command explanation.

• 🖺 No-brainer RAG (Retrieval Augmented Generation)

• 🧠 Experiment any model. (Llama3, Phi4, DeepSeek-R1, Qwen) with parameters of your choice.

Installation

Installation can be done in two ways;

• Installation script (recommended)
• Go Install

Installation Script

Installation script is the recommended way to install tlm. It will recognize the which platform and architecture to download and will execute install command for you.

Linux and macOS;

Download and execute the installation script by using the following command;

curl -fsSL https://raw.githubusercontent.com/yusufcanb/tlm/1.2/install.sh | sudo -E bash

Windows (Powershell 5.5 or higher)

Download and execute the installation script by using the following command;

Invoke-RestMethod -Uri https://raw.githubusercontent.com/yusufcanb/tlm/1.2/install.ps1 | Invoke-Expression

Go Install

If you have Go 1.22 or higher installed on your system, you can easily use the following command to install tlm;

go install github.com/yusufcanb/tlm@1.2

You're ready! Check installation by using the following command;

tlm

Usage

$ tlm
NAME:
   tlm - terminal copilot, powered by open-source models.

USAGE:
   tlm suggest "<prompt>"
   tlm s --model=qwen2.5-coder:1.5b --style=stable "<prompt>"

   tlm explain "<command>" # explain a command
   tlm e --model=llama3.2:1b --style=balanced "<command>" # explain a command with a overrided model

   tlm ask "<prompt>" # ask a question
   tlm ask --context . --include *.md "<prompt>" # ask a question with a context

VERSION:
   1.2

COMMANDS:
   ask, a      Asks a question (beta)
   suggest, s  Suggests a command.
   explain, e  Explains a command.
   config, c   Configures language model, style and shell
   version, v  Prints tlm version.
   help, h     Shows a list of commands or help for one command

GLOBAL OPTIONS:
   --help, -h     show help
   --version, -v  print the version

Ask - Ask something with or without context

Ask a question with context. Here is an example question with a context of this repositories Go files under ask package.

$ tlm ask --help
NAME:
   tlm ask - Asks a question (beta)

USAGE:
   tlm ask "<prompt>" # ask a question
   tlm ask --context . --include *.md "<prompt>" # ask a question with a context

OPTIONS:
   --context value, -c value                                context directory path
   --include value, -i value [ --include value, -i value ]  include patterns. e.g. --include=*.txt or --include=*.txt,*.md        
   --exclude value, -e value [ --exclude value, -e value ]  exclude patterns. e.g. --exclude=**/*_test.go or --exclude=*.pyc,*.pyd
   --interactive, --it                                      enable interactive chat mode (default: false)
   --model value, -m value                                  override the model for command suggestion. (default: qwen2 5-coder:3b)
   --help, -h                                               show help

Suggest - Get Command by Prompt

$ tlm suggest --help
NAME:
   tlm suggest - Suggests a command.

USAGE:
   tlm suggest <prompt>
   tlm suggest --model=llama3.2:1b <prompt>
   tlm suggest --model=llama3.2:1b --style=<stable|balanced|creative> <prompt>

DESCRIPTION:
   suggests a command for given prompt.

COMMANDS:
   help, h  Shows a list of commands or help for one command

OPTIONS:
   --model value, -m value  override the model for command suggestion. (default: qwen2.5-coder:3b)
   --style value, -s value  override the style for command suggestion. (default: balanced)        
   --help, -h               show help

Explain - Explain a Command

$ tlm explain --help
NAME:
   tlm explain - Explains a command.

USAGE:
   tlm explain <command>
   tlm explain --model=llama3.2:1b <command>
   tlm explain --model=llama3.2:1b --style=<stable|balanced|creative> <command>

DESCRIPTION:
   explains given shell command.

COMMANDS:
   help, h  Shows a list of commands or help for one command

OPTIONS:
   --model value, -m value  override the model for command suggestion. (default: qwen2.5-coder:3b)
   --style value, -s value  override the style for command suggestion. (default: balanced)        
   --help, -h               show help

Uninstall

On Linux and macOS;

rm /usr/local/bin/tlm
rm ~/.tlm.yml

On Windows;

Remove-Item -Recurse -Force "C:\Users\$env:USERNAME\AppData\Local\Programs\tlm"
Remove-Item -Force "$HOME\.tlm.yml"

Contributors