tninja/aider.el

AI assisted programming in Emacs with Aider

What it is

aider.el is an Emacs package that provides a user interface for Aider, a popular terminal-based AI pair-programming tool. It brings AI-assisted coding (similar to Cursor) into the Emacs editor.

Core idea

It acts as a lightweight wrapper around the Aider command-line tool. Instead of leaving Emacs to use Aider in a terminal, you can run and control Aider sessions directly inside Emacs, with menus, keybindings, and buffer management.

Key features

  • Session management: Starts Aider sessions tied to specific Git repositories, allowing multiple simultaneous sessions for different projects.
  • Context control: Easily add files to the AI's context—the current file, files marked in a file manager (dired), or entire modules/directories.
  • Coding workflows:
    • Implement TODO comments in code.
    • Refactor code using techniques from Martin Fowler's book.
    • Generate unit tests for functions or files.
    • Ask questions about selected code or functions.
    • Pull and review code diffs (from staged files, branches, or commits).
  • Prompt organization: A dedicated "prompt file" per repository where you can write, organize, and send multi-line prompts to the AI session.
  • Agile support: Includes tools for TDD (test-driven development), refactoring, and handling legacy code.
  • Integration: Works with Magit (Git interface), Flycheck (error checking), and Helm (fuzzy search).

How it works

  • It requires the Aider CLI to be installed separately.
  • Configuration is passed directly to the Aider CLI, so it shares the same settings and model options (e.g., Claude, GPT) as the terminal version.
  • It uses a "transient" menu (triggered by a keybinding like C-c a) to show all available commands.
  • It supports various AI models, including local ones via LiteLLM.

Requirements

  • Emacs 26.1 or newer.
  • The aider CLI installed.
  • Emacs packages: transient, magit, and markdown-mode.
  • An API key for an AI model (e.g., Anthropic or OpenAI).

In short, it's a productivity tool that makes AI pair-programming a first-class citizen inside Emacs, with a focus on code review, refactoring, and structured workflows.

Related

  • Project
    aovestdipaperino/tokensave

    tokensave is a Rust‑based local server that builds a semantic knowledge graph of a codebase, letting AI coding agents query symbols, relationships, and snippets in a single call. By replacing costly grep/glob scans, it cuts token usage, supports multi‑branch indexing, offers crash‑resilient syncing, and integrates with dozens of LLM‑powered coding tools.

  • Project
    WEIFENG2333/phistory

    Phistory archives and visualises system prompts of many AI coding‑assistant CLIs, letting users compare how agents’ hidden instructions change over time.

  • Project
    khromov/ai-digest

    ai-digest is a Node‑based CLI (and library) that bundles a project's source files into a single Markdown document, applying ignore/minify rules and optional whitespace stripping, so the whole codebase can be fed to Claude or custom ChatGPT agents for analysis.

  • Project
    numpy/numpy-tutorials

    A repository of official, high-quality NumPy tutorials for learning and teaching, separate from the core NumPy code.