Aider
Free tierAI pair programming in your terminal — code smarter with LLMs
Free·Technical·Powered by Anthropic, OpenAI, DeepSeek, and others·Open source
Key strengths
Supports 100+ languages and connects to virtually any LLM (cloud or local)Whole-codebase mapping enables accurate, context-aware edits in large projectsNative Git integration with automatic, meaningful commit messagesRuns entirely in the terminal — no cloud service or account requiredVoice-to-code, image/webpage context, and IDE comment-driven workflows
Completely free
Self-hostable
No ratings yet
Technical Setup & Key Parameters
Installation
python -m pip install aider-install && aider-install
Aider supports Python 3.9+. It uses LiteLLM under the hood, giving it compatibility with virtually any LLM provider or local inference server.
Model Selection
aider --model <model-name> --api-key <provider>=<key>
Recommended models: claude-3-7-sonnet, deepseek, gpt-4o, o3-mini. Local models can be targeted via Ollama or any OpenAI-compatible endpoint.
Key CLI Flags
| Flag | Description |
|---|---|
--model | Select the LLM to use |
--api-key | Pass provider API keys inline |
--auto-commits / --no-auto-commits | Toggle automatic Git commits |
--lint | Run linter after each change |
--test-cmd | Shell command to run your test suite |
--voice | Enable voice-to-code input |
--read | Add files/URLs as read-only context |
Repo Map
Aider uses tree-sitter to parse your codebase into a symbol map that is injected into the LLM context window, allowing accurate multi-file edits without requiring you to manually specify every relevant file.
IDE Integration
Aider can be invoked from any IDE terminal. It also supports comment-driven mode: add a special comment in your source code and Aider will pick it up and implement the requested change.