aider capabilities

Aider's architect mode uses two separate LLMs in sequence: one model proposes the solution (architect), and a second model translates that proposal into precise file edits (editor).

After every AI-driven file edit, Aider automatically commits the changes to git with a descriptive commit message, making it easy to track, review, or undo AI changes.

Aider connects to virtually any LLM — OpenAI, Anthropic, Google Gemini, DeepSeek, Groq, Ollama, local models, and any OpenAI-compatible API — using the user's own API keys.

Aider can load one or more markdown files containing coding conventions, style rules, or project guidelines that persist as read-only context across the entire session.

Aider provides a rich set of slash commands within the chat REPL for controlling files, models, git, tests, and session state without leaving the terminal.

Aider publishes a public benchmark leaderboard that ranks LLMs on their ability to write and edit code, helping users choose the best model for pair programming.

Aider automatically lints edited files after each AI change and can run a test suite, looping to fix any errors reported by non-zero exit codes.

Aider can propose and apply coordinated edits across multiple files in a single response, using structured diff or whole-file formats to make precise changes.

Aider itself is free, open-source software (Apache-2.0 license) with no subscription or seat fee. Users bring their own LLM API keys; LLM usage costs are paid directly to the model provider.

Aider builds a concise structural map of the entire git repository — listing files with their key classes, functions, and type signatures — so the LLM understands the full codebase without loading every file.

Aider runs entirely in the terminal as a CLI tool, letting developers pair-program with LLMs to build new projects or modify existing codebases via conversational prompts.

Aider lets developers speak their coding requests into the terminal; the audio is transcribed and submitted as a chat message, so the AI can then edit files accordingly.

Aider can monitor all repo files for special AI comment markers, triggering code changes or Q&A directly from annotations added in any editor — no terminal interaction needed after launch.

Aider can fetch and include the text content of any web page as context in the chat, making it easy to pull in live documentation, API references, or any URL the LLM needs.