Aider Review
Open-source AI pair programmer that edits your codebase from the terminal
Is this your product? Claim this page · Request a change
Looking for a Aider alternative? See our ranked comparison.→What is Aider?
Aider is a command-line tool that edits your codebase directly based on instructions you type in chat. It reads your repo, builds a map of the code, sends relevant files to an LLM, and applies the diffs the model proposes, committing each change to git automatically so you can review or roll back with normal git commands.
It isn't a company product. It's an open-source project you install with pip and run against whatever LLM API key you already have. There's no account, no dashboard, and no bundled model, so your experience and your bill both depend on which model you choose to pair it with.
Aider screenshots
Who it's for
- ✓ Developers who already pay for an LLM API key and want a terminal-native editor loop instead of a chat window
- ✓ Teams that want full control over which model they use, including cheaper or local models
- ✓ Anyone who wants every AI edit to land as a normal git commit they can diff and revert
Who should look elsewhere
- ✗ Developers who want a single bill and a polished GUI instead of managing their own API key and provider costs
- ✗ Teams that want an included model, official support, or an SLA to call when something breaks
- ✗ People who want the highest raw coding benchmark scores out of the box without picking and tuning a model themselves
Pros
- + Free and open source, no subscription or seat cost
- + Works with almost any LLM, including Claude, GPT, DeepSeek, Gemini, and local models via Ollama
- + Every change is a git commit with a real message, so undoing an edit is just `git revert`
- + Repo map feature helps it stay useful in large codebases without dumping the whole tree into context
- + Actively maintained on GitHub, with commits adding new model support arriving every few weeks even when packaged releases lag behind
Cons
- – No pricing means no cost predictability. Your bill scales with token usage and the model you pick, not a flat fee
- – No official support channel. Help comes from GitHub issues and community Discord, not a vendor
- – Terminal-only workflow with no built-in GUI, which is a bigger adjustment for developers coming from an IDE-integrated tool
- – Scores lower on SWE-bench Verified (about 60%) than Claude Code (about 81%) and Codex (about 80%) in third-party benchmarks, so out-of-the-box task completion can lag
Aider pricing
What you pay for
Aider has no price tag of its own. You pay your LLM provider for tokens, and the bill depends entirely on which model you point Aider at and how much code you send it. There's no tier to pick and no seat to buy. That also means there's no vendor to call if a bill surprises you.
You pay for what you consume rather than a per-seat fee, so cost scales with usage.
| Plan | Price | Highlights |
|---|---|---|
| Pricing is quote-only. Contact the vendor for a quote. | ||
Aider itself is free and open source (Apache-2.0). There is no subscription or account. You install it with pip and bring your own API key for whichever LLM provider you pick (Anthropic, OpenAI, DeepSeek, Gemini, local models via Ollama, and more). What you pay is whatever that provider charges for tokens, billed directly by the provider, not by Aider. There is no vendor pricing page to quote a number from.
Pricing verified July 22, 2026 · source
How Aider's pricing compares
Aider next to its closest alternatives on entry price, billing, and whether pricing is public.
| Tool | Starting price | Billing | Free option | Pricing disclosed |
|---|---|---|---|---|
| Aider | Free tier + custom | usage-based | Yes | Not disclosed |
| Cursor | $20/seat/mo | usage-based | Yes | Partly public |
| GitHub Copilot | $10/seat/mo | usage-based | Yes | Public |
| Windsurf | $20/seat/mo | per-seat | Yes | Partly public |
| Tabnine | $39/seat/mo | per-seat | No | Partly public |
| Sourcegraph Cody | $16000/yr | quote-only | No | Partly public |
Is Aider still actively developed?
Last significant update: May 2026. The last tagged PyPI release is still 0.86.2 from February 2026, itself just a version bump with no feature notes. Development kept going on GitHub past that point, most recently adding Claude Opus 4.5 model aliases, part of a run of commits through Q2 2026 that added support for GPT-5.5 and newer Claude models as they shipped, none of which has been packaged into a new release yet.
Top Aider alternatives
Aider FAQ
Is Aider free?+
Aider itself is free and open source. You still pay your LLM provider (Anthropic, OpenAI, DeepSeek, and so on) for the API tokens it uses, so your actual cost depends on which model you choose and how much code you send it.
Which LLM should I use with Aider?+
Aider works with almost any LLM, including local models through Ollama. The project's docs recommend strong coding models like Claude and GPT-4 class models for the best results, but cheaper models will cost less per feature at some quality tradeoff.
Does Aider have a company behind it with support and an SLA?+
No. Aider is maintained as an open-source project under the Apache-2.0 license, with no company, subscription, or official SLA. Support comes from GitHub issues and the community.