AItoComplete

AItoComplete is an Emacs package that provides a convenient interface to Ollama, allowing you to access LLM capabilities directly from your Emacs workflow.

Prerequisites

• Emacs 26.1 or higher
• Ollama installed and running locally
• curl command-line tool

Installation

Manual Installation

1. Download aitocomplete.el to a directory in your Emacs load path.
2. Add the following to your init.el:

(require 'aitocomplete)

Using straight.el and use-package

(use-package aitocomplete
  :straight (:host github :repo "hipml/aitocomplete")
  :bind (("C-c s" . aitocomplete-send-region)
         ("C-c a" . aitocomplete-menu)))

Using MELPA (once it's available)

(use-package aitocomplete
  :ensure t)

Configuration

The package works out of the box, but you can customize it to your preferences:

;; Set your preferred model
(setq aitocomplete-model "llama3.2")  ; Default model

;; Customize buffer name
(setq aitocomplete-chat-buffer "*My AI Assistant*")

;; Set the number of columns in the menu
(setq aitocomplete-menu-columns 4)

;; Custom keybinding for the menu
(global-set-key (kbd "C-c a") #'aitocomplete-menu)

Usage

Quick Start

1. Start the Ollama server on your local machine
2. Select a region of text in any buffer
3. Press C-c s to send the region to the AI
4. View the AI's response in the chat buffer

Main Commands

• M-x aitocomplete-menu: Display the main menu with all options
• M-x aitocomplete-send-region: Send the selected region to the AI
• M-x aitocomplete-chat: Open or switch to the AI chat buffer
• M-x aitocomplete-test-response: Test the AI connection with a simple prompt

In the Chat Buffer

• Type your message
• Press C-c C-c to send it to the AI
• Press ? to display the menu

Default Keybindings

Keybinding	Command	Description
C-c s	aitocomplete-send-region	Send selected text to AI
C-c C-c	aitocomplete-send-message	(In chat buffer) Send message
?	aitocomplete-menu	(In chat buffer) Display menu

Menu Options

Press M-x aitocomplete-menu or ? in the chat buffer to access these options:

• [o] Open chat buffer
• [s] Send region
• [m] Change model
• [t] Test response
• [q] Quit menu

Supported Models

By default, the package supports whatever models you have available in your local Ollama installation. You can check available models by opening the menu.

Troubleshooting

Ollama Server Not Running

If you see "NOT RUNNING" in the menu, make sure:

1. Ollama is installed on your system
2. The Ollama server is running
3. It's accessible at http://localhost:11434

Debug Messages

Enable debug messages to help troubleshoot:

(setq debug-on-error t)

Contributing

Contributions are welcome! Feel free to open issues or submit pull requests on GitHub.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

AItoComplete was inspired by similar projects that integrate AI capabilities into Emacs.