Aider is a free, open-source AI pair programmer that runs in your terminal and commits every code change to git automatically.
Aider is an open-source AI pair programmer that runs directly in your terminal instead of inside an IDE, building a map of your entire codebase so it works well even on large, unfamiliar projects. You point it at a git repository, add the files you want to work on, describe the change in plain English, and Aider edits the code, runs your linters and tests, and commits every successful change automatically with a descriptive git message. Because it stays this close to git and the command line, Aider gives developers building production AI systems a fast, auditable way to make multi-file changes without leaving the workflow they already trust.
Aider works from inside a normal git repository on your machine. You add the specific files relevant to a task to the chat session, then describe what you want changed using plain English, comments in the code, or even your voice.
Repo mapping: Aider scans your entire codebase and builds a lightweight map of classes, functions, and signatures, so it understands how files depend on each other even when they are not in the chat.
File edits: Once you describe a change, Aider writes the actual code, working across multiple files in a single coordinated edit when the task requires it.
Git commits: Every successful edit is staged and committed to git automatically with a descriptive, model-generated commit message, so you can diff, revert, or branch from any point with tools you already know.
Lint and test loop: Aider can run your linter and test suite after every change. If something breaks, it feeds the error back to the model and asks for a fix before moving on.
Flexible model backend: Aider connects to Claude, GPT-4o, o3-mini, DeepSeek, Gemini, or local models through Ollama, so you are never locked into one provider.
Feature additions on an existing codebase: Describe a new feature in plain English and let Aider implement it across the relevant files, using the repo map to understand how the pieces connect.
Bug fixes with a test-driven loop: Point Aider at a failing test, let it read the stack trace, and have it patch the code and rerun the suite until the test passes.
Full-stack app prototypes: Start a new project from a single prompt and let Aider scaffold the front end, back end, and configuration files needed to run it.
Legacy code refactors: Use Aider’s git integration to make large, multi-file refactors safely, since every change lands as its own reversible commit.
Local, offline coding workflows: Connect Aider to a local model through Ollama and get AI pair programming with zero API cost and no code leaving your machine.
Voice-driven coding sessions: Describe a feature or fix out loud and let Aider translate the request directly into code changes.
Aider itself is free and open-source under the Apache 2.0 license. You only pay for the LLM API calls you make through providers like Anthropic or OpenAI, or use it for free entirely with local models through Ollama. Typical usage costs range from a few cents to a few dollars per coding session depending on the model.
Aider runs in your terminal and connects to an LLM of your choice. You add files to the chat, describe the change you want in plain English, and Aider edits the code directly, then commits the change to git with an auto-generated message. It also builds a map of your codebase so the model understands unfamiliar files.
Aider works with almost any LLM, including Claude, GPT-4o and o3-mini from OpenAI, DeepSeek R1 and Chat V3, Gemini, and local models run through Ollama. It performs best with frontier coding models like Claude Sonnet, but you can switch providers at any time without changing your workflow.
Claude Code is Anthropic's terminal-based AI coding agent that reads your codebase, edits files, runs commands, and opens pull requests.
Coding AgentCodeium's AI-native code editor, built around the Cascade agent, that Cognition rebranded to Devin Desktop in June 2026 after acquiring the company.
Coding AgentCursor is an AI-first code editor that ships with deeply integrated copilot, multi-file edits, and an agent mode.
Coding AgentUpdates from the AI world — what shipped, what we’re using in production, and what’s worth your attention. Two emails a month, no spam.
