Run Aider with a Local Model on Your Mac

Aider is the terminal pair-programmer that edits files in place. Point it at a model served on your own Mac and your prompts and code stay on the Mac, with no per-token bill.

Aider reads the files you add, sends them to a model with your instruction, and applies the edits it gets back. It reaches models through LiteLLM, so any OpenAI-compatible endpoint works, including Rapid-MLX on your Mac.

• Your code stays local. Aider sends files to localhost, not to a cloud model.
• No meter. Edit all day for the cost of electricity.
• Forgiving with small models. Aider's edit blocks are plain text, not tool-call JSON, so small local models handle it well.

The whole thing

1. Install Rapid-MLX and serve a model.
2. Export OPENAI_API_BASE, OPENAI_API_KEY and AIDER_MODEL.
3. aider <file>.

1 · Install Rapid-MLX and Aider

curl -fsSL https://rapidmlx.com/install.sh | bash
uv tool install --python 3.12 aider-chat

uv tool puts Aider in its own environment (pip install aider-chat works too). Rapid-MLX is also in Homebrew core: brew install rapid-mlx.

2 · Serve a model

rapid-mlx serve qwen3.5-4b-4bit

That's the engine's pick for a 16 GB Mac and the model we verified with. rapid-mlx recipe prints the pick for your memory; Best local LLMs by Mac shows measured speeds per chip. Keep the server running.

3 · Point Aider at it

rapid-mlx agents aider --setup prints the three variables for the model you're serving. Export them in the shell you'll run Aider from:

export OPENAI_API_BASE=http://localhost:8000/v1
export OPENAI_API_KEY=not-needed
export AIDER_MODEL=openai/qwen3.5-4b-4bit

4 · Run it

Open Aider on a file and describe the change, or pass it as a one-shot message:

# inside your project folder
aider --message "Add a function add(a, b) to hello.py that returns a + b." hello.py

From our run:

hello.py
@@ -1,2 +1,5 @@
 def greet(name):
     return "Hello, " + name
+
+def add(a, b):
+    return a + b

Tokens: 757 sent, 44 received.
Applied edit to hello.py

Frequently asked questions

Can Aider use a local model on a Mac?

Yes. Set OPENAI_API_BASE=http://localhost:8000/v1 and an openai/-prefixed model name, serve the model with Rapid-MLX, and Aider edits your files with the model on your Mac.

Why does the model name need the openai/ prefix?

Aider talks to models through LiteLLM, and the openai/ prefix tells LiteLLM to use the OpenAI-compatible route with your base URL. Without it, LiteLLM tries to match a cloud provider.

Which local model is best for Aider?

Aider is easy on small models, so even qwen3.5-4b-4bit handles single-file edits. For larger changes, use the biggest pick rapid-mlx recipe suggests, such as qwen3.8-27b-4bit from 32 GB.

Where to go next

• Aider reference: model naming, edit formats and recommended aliases.
• Best local LLMs by Mac: measured speed for your chip and memory.
• Rapid-MLX vs mlx-lm: the same weights, measured task by task.
• Other agents: Claude Code, Codex CLI, OpenCode and more.

curl -fsSL https://rapidmlx.com/install.sh | bash