Aider + Ollama Setup

Configure Aider to use local Ollama models for fast, private AI-powered coding without cloud API costs.

What This Does

This skill sets up Aider with Qwen2.5-Coder models running locally via Ollama. You get:

Prerequisites

1. **Install Ollama**: [ollama.com](https://ollama.com)

2. **Pull models**:

```bash

ollama pull qwen2.5-coder:7b

ollama pull qwen2.5-coder:3b

```

3. **Install Aider**:

```bash

pip install aider-chat

```

Instructions

Step 1: Create Configuration File

Create `~/.aider.conf.yml` with the following content:

```yaml

Default model - Qwen2.5-Coder 7B (Fast & Smart!)

model: ollama/qwen2.5-coder:7b

Weak model for simpler tasks (faster)

weak-model: ollama/qwen2.5-coder:3b

Edit format - how aider makes changes

edit-format: diff

Git integration

auto-commits: true

dirty-commits: true

auto-lint: true

Output preferences

pretty: true

stream: true

dark-mode: true

Show git diffs when reviewing changes

show-diffs: true

Map settings (code context)

map-tokens: 2048

map-refresh: auto

Chat history

restore-chat-history: true

Don't show model warnings for local models

show-model-warnings: false

```

Step 2: Verify Ollama is Running

```bash

Check Ollama status

ollama list

Should show qwen2.5-coder models

```

Step 3: Test Aider

```bash

Start aider in your project

aider

Try a simple command

/help

```

Step 4: Start Coding

```bash

Add files to edit

/add src/main.py

Make changes

> Add error handling to the main function

Review changes before committing

/diff

Commit when ready

/commit

```

Configuration Details

Model Strategy

Git Integration

Context Management

UX Settings

Custom Ollama Connection

If Ollama runs on a different host or port:

```yaml

Add to ~/.aider.conf.yml

openai-api-base: http://your-host:11434/v1

```

Usage Tips

1. **Start with context**: Add relevant files with `/add` before making requests

2. **Review before committing**: Use `/diff` to see changes before they're committed

3. **Use architect mode**: For big changes, use `/architect` to plan before coding

4. **Drop unnecessary files**: Use `/drop` to remove files from context and save tokens

5. **Undo mistakes**: Use `/undo` to revert the last aider commit

Why Qwen2.5-Coder?

Troubleshooting

**Ollama not connecting?**

```bash

Check Ollama is running

curl http://localhost:11434/api/tags

```

**Model not found?**

```bash

Verify models are downloaded

ollama list

ollama pull qwen2.5-coder:7b

```

**Slow performance?**

Alternative Models

Want to try different models? Edit `~/.aider.conf.yml`:

```yaml

DeepSeek Coder (33B - more powerful, slower)

model: ollama/deepseek-coder:33b

CodeLlama (13B - good middle ground)

model: ollama/codellama:13b

Llama 3 (general purpose)

model: ollama/llama3:8b

```

Then pull the model: `ollama pull model-name`

Documentation