Cursor
AI-native code editor that runs autonomous agents to build, test, and review code
Open-source AI pair programming in your terminal — bring your own LLM key and edit your git repo
Aider is a free, open-source AI pair programmer that runs in your terminal and edits files directly in your local git repo, auto-committing each change. The tool costs $0; you bring your own LLM API key, so the only cost is your model usage — or nothing at all if you run a local model. It is model-agnostic across Claude, GPT, DeepSeek, and others, and on its polyglot leaderboard the top models score around 88%. Best for terminal-comfortable developers.
Aider is a free, open-source AI pair-programming tool that runs in your terminal. You describe what you want changed, and Aider edits the relevant files directly in your local git repository, committing each change with a generated commit message so the work stays reviewable and reversible. Rather than bundling its own model and subscription, Aider is model-agnostic: you bring your own API key and point it at whatever LLM you prefer — Claude, GPT, DeepSeek, Gemini, or a local model run through Ollama. The software itself costs nothing; your only spend is the model usage you choose.
That design makes cost and quality your decision, not the vendor’s. Aider builds a “repo map” to give the model context on large codebases, supports more than 100 languages, and adds quality-of-life features like watch mode (responding to inline comments from your editor), voice-to-code, image and web-page inputs, and automatic lint-and-test fixing. The project also publishes a closely watched polyglot leaderboard, where the strongest models score around 88%, giving an honest, public reference for which model to pair with Aider. Claude’s coding models are among the best to run it with.
Aider is for developers who are comfortable in the terminal and want maximum control over which model they use and what it costs. It rewards people who value git-native, reviewable changes and dislike being locked into a single vendor’s model or subscription. Developers who prefer a polished graphical editor may find tools like Cursor or Windsurf a gentler start.
Starting price: $0 · Free tier: yes · Model: open-source
Price history tracked from June 2026
| Plan | Price | Includes |
|---|---|---|
| Aider (open source) | Free | MIT-licensed; install via pip, no account or subscription · Bring your own LLM API key — pay only your model provider · All features included; no paywalled tiers · Runs free end-to-end with a local model via Ollama |
| Pros | Cons |
|---|---|
| Genuinely free and open source — no vendor lock-in to any model or subscription | Not free in practice — you pay real API costs that add up on large frontier models |
| Best-in-class model flexibility; you control cost by choosing cheaper or local models | Terminal-only with no GUI, a steeper learning curve than editor tools like Cursor |
| Deep, automatic git integration makes every AI change reviewable and reversible | Requires you to obtain and manage API keys and understand model trade-offs |
| Maintains a respected public leaderboard, so model-choice guidance is transparent | Output quality depends entirely on the model you supply — a weak model gives weak results |
AI-native code editor that runs autonomous agents to build, test, and review code
Open-source autonomous coding agent for VS Code that edits files, runs commands, and brings your own model
AI coding assistant with inline completions, chat, and autonomous agent mode across major IDEs
Agentic AI IDE with Cascade that writes, edits and runs code autonomously inside your editor
Anthropic's AI assistant built for long-context analysis, coding, and safe reasoning
The Aider software is free and open source (MIT-licensed) — you install it with pip and there is no account or subscription. The real cost is whatever you spend on your chosen LLM's API usage. If you run a local model through Ollama, it can be free end-to-end.
Aider is model-agnostic and works with Claude, GPT, DeepSeek, Gemini, and local models. Claude's coding models are among the strongest to pair it with. Aider publishes a polyglot leaderboard where top models score around 88%, which is the best guide to current model choice.
Yes — git is central to how Aider works. It commits each change automatically with a descriptive message, treating git as the source of truth. That makes every AI edit easy to review, diff, and roll back if needed.
Aider runs entirely in the terminal and lets you bring any model via your own API key, whereas Cursor and GitHub Copilot are GUI editor experiences with bundled subscriptions. Aider trades a polished UI for model flexibility and direct git control.
Yes. Aider is a terminal application, so you should be comfortable with the command line, installing Python packages, and managing API keys. Developers who prefer a graphical editor may find tools like Cursor or Windsurf easier to start with.
Yes. Aider builds a repo map to give the model context about a large project's structure, and it supports over 100 programming languages. Performance and cost still depend on the model you choose and how much context each request needs.