Overview

A personal LLM command-line toolkit. Pipe text into named tasks and get processed output. Supports both local models via Ollama and cloud APIs (OpenAI and Anthropic). On macOS, integrates with the clipboard and system notifications.

Features

• Named tasks managed in a single configuration file with user-defined instructions
• Translation with automatic language detection via whichlang-cli
• Streaming output from all providers
• Works as a standalone CLI or sourced as a library

Requirements and Dependencies

Required

Dependency	Version	Notes
bash	v3.2 or later	Compatible with the version bundled on macOS
curl	—	HTTP requests
jq	—	JSON processing

Optional

Dependency	Purpose
whichlang-cli	Automatic language detection for translation (supports 16 languages)
ollama	Running local models

Installation

Run the interactive installer:

bash -c "$(curl -fsSL https://raw.githubusercontent.com/rinodrops/myllm-cli/main/install.sh)"

What the installer does:

1. Checks for required dependencies (jq, curl) and offers to install any that are missing
2. Installs myllm to /usr/local/bin or ~/.local/bin
3. Optionally installs whichlang-cli
4. Creates ~/.config/myllm/config.toml from a default template

Quick Start

1. Configure a Provider

First, configure a provider in ~/.config/myllm/config.toml. If you use Ollama, the defaults work out of the box:

[general]
default_provider = "ollama"

[providers.ollama]
base_url = "http://localhost:11434"
default_model = "llama3.2"

To use a cloud API, see the Configuration Reference.

2. Define a Task

[tasks.polish]
name = "Polish"
instruction = '''
Improve the clarity and correctness of the provided text.
Output only the revised text, no explanation.
Use the same language as the original text.
'''

3. Run

# Pipe text
echo "draft email" | myllm polish

# Pass as an argument
myllm polish "draft email"

# List available tasks
myllm list

For detailed command specifications, see the Command Reference.