Get Started
← Own your AI · Mission 13 of 31

An AI pair programmer in your terminal that edits your code and commits it, using a free local model

Fix a bug by typing one sentence, watch the changes appear, and throw away anything you don't like with /undo. Aider edits your project and saves every change as a git commit, using a model on your own laptop, so your code never leaves your machine.

40 minIntermediateTools: FreeStep 1 free
Tested 9 Oct 2026 · Aider 0.86.2, Ollama 0.40.1, Qwen2.5-Coder 3B, uv 0.11.9, git 2.52, Windows 11 (16GB, no graphics card)
Aider installed and talking to a coding model that runs on your own laptop
A real bug fixed from one sentence, checked with a test, and saved as a git commit
A feature added across two files, and /undo to throw away a change you don't like
  • A short checklist for getting good results from a small model, and a bigger-brain option for when you need more
Before you start
  • Ollama installed. Not done that yet? Do Run a ChatGPT-style AI on your own laptop first (15 minutes).
  • git installed (winget install --id Git.Git -e on Windows; Macs offer to install it the first time you type git). New to git? Git without memorising covers the basics.
  • About 2.5GB of free disk space: 1.9GB for the model and about 480MB for Aider
  • 16GB of RAM recommended
01/06Download a small coding model0 done
STEP 01 of 06~5 min

Download a small coding model

Aider is only the messenger. The model running in Ollama is what actually writes the code, so pick one trained for code.

  1. If you've done the Ollama lesson, Ollama is already running. If not, install it first:

    Windows (skip if you have Ollama)
    winget install --id Ollama.Ollama -e

  2. Now download Qwen2.5-Coder 3B, a small model made for writing and fixing code:

    Terminal
    ollama pull qwen2.5-coder:3b

  3. It's a 1.9GB download, a few minutes on a typical connection. Check it's there:

    Terminal
    ollama list
    Output
    NAME                  ID              SIZE      MODIFIED
qwen2.5-coder:3b      f72c60cabf62    1.9 GB    4 seconds ago
...
    TipWhy start small? Once loaded, the 3B model uses about 2.4GB of memory (ollama ps shows this), so it runs on most laptops. The 7B version is smarter but a 4.7GB download that needs a lot more memory. Step 6 shows how to switch to it once everything works.
✓
Checkpointollama list shows qwen2.5-coder:3b.
If it breaks'could not connect to a running Ollama instance'
'could not connect to a running Ollama instance'
Ollama isn't running in the background. Open Ollama from the Start menu (or Applications on a Mac) and try again.
Still stuck? Ask the community
Next mission · Own your AIRun a private AI chatbot in a browser tab, with one click
→