Skills matrix
| Skill | Status | Delivery | Maturity | Evidence |
|---|---|---|---|---|
| Code generation | Partial | Model-dependent | Model-limited | Not publicly documented |
| Refactoring | Partial | Model-dependent | Model-limited | Not publicly documented |
| Multi-file reasoning | Partial | Model-dependent | Model-limited | Not publicly documented |
| Test generation | Partial | Model-dependent | Model-limited | Not publicly documented |
| Static analysis support | Partial | Model-dependent | Model-limited | Not publicly documented |
| Codebase indexing | Partial | Model-dependent | Model-limited | Not publicly documented |
| Semantic retrieval | Partial | Model-dependent | Model-limited | Not publicly documented |
| Memory retention across sessions | Partial | Model-dependent | Model-limited | Not publicly documented |
| Context injection control | Partial | Model-dependent | Model-limited | Not publicly documented |
| Inline editing | Partial | Model-dependent | Model-limited | Not publicly documented |
| Chat-first interaction | Partial | Model-dependent | Model-limited | Not publicly documented |
| File diff preview | Partial | Model-dependent | Model-limited | Not publicly documented |
| Git integration | Partial | Model-dependent | Model-limited | Not publicly documented |
| Terminal / command execution | Partial | Model-dependent | Model-limited | Not publicly documented |
| Model switching | Partial | Model-dependent | Model-limited | Not publicly documented |
| Multi-model orchestration | Partial | Model-dependent | Model-limited | Not publicly documented |
| Prompt augmentation layer | Partial | Model-dependent | Model-limited | Not publicly documented |
| Agent loop execution | Partial | Model-dependent | Model-limited | Not publicly documented |
| Local model support | Partial | Model-dependent | Model-limited | Not publicly documented |
| Offline capability | Partial | Model-dependent | Model-limited | Not publicly documented |
| Enterprise policy control | Partial | Model-dependent | Model-limited | Not publicly documented |
Why it matters
- One of the highest-rated AI coding tools on its benchmark leaderboard — particularly strong for multi-file refactoring.
- Automatic git commit after every change: creates a clean, reviewable history of all AI edits.
- Works in your existing terminal and editor — no IDE switch, no subscription lock-in, bring your own LLM.
- Repository map provides accurate codebase context for much larger projects than naive file-inclusion approaches.
Key capabilities
- Multi-file editing: Edit multiple files in a single request — Aider coordinates changes across your entire codebase.
- Automatic git commits: Every edit is committed to git with an AI-generated descriptive commit message.
- Repository map: Compact codebase graph (imports, symbols, classes) keeps large repos within model context limits.
- LLM flexibility: Works with Anthropic Claude, OpenAI GPT-4o, Google Gemini, DeepSeek, Groq, and local Ollama models.
- In-chat commands:
/add <file>to include files,/runto execute code,/testto run tests,/undoto revert last commit. - Code review mode:
--reviewmode explains changes without applying them — good for reviewing AI suggestions. - Voice input:
--voicemode for spoken instructions (experimental, requires microphone). - Watch mode:
--watchmonitors for AI comments in code (e.g.,# AI: refactor this) and applies them.
Technical notes
- License: Apache 2.0 — fully open source at github.com/paul-gauthier/aider
- Install:
pip install aider-chat; requires Python 3.8+ - LLMs: Anthropic API, OpenAI API, Google Gemini API, Groq, Ollama (local), any OpenAI-compatible endpoint
- Context: Repository map + explicitly added files; configures automatically based on project size
- Benchmark: aider.chat/docs/leaderboards — tracks completion rates across 133+ coding exercises
- Author: Created by Paul Gauthier; active open source project with weekly releases
- Shell integration: Works in any terminal; no VS Code dependency
Ideal for
- Developers who prefer command-line tools and want AI code editing without leaving the terminal.
- Engineers doing large-scale refactors or cross-file changes that require consistent multi-file edits.
- Power users who want precise control over LLM choice, context, and commit strategy.
Not ideal for
- Beginners unfamiliar with git or the command line — Cursor or Lovable have more accessible UIs.
- Teams who prefer visual code review of AI changes before they're committed.
- Projects without version control — Aider's workflow centers on git; non-git directories require
--no-git.
See also
- Claude Code — Anthropic's official CLI coding agent, similar terminal-first approach.
- Cline — VS Code extension-based AI agent for automated multi-step coding tasks.
- Cursor — AI-native IDE with visual chat, inline editing, and multi-file agents.