Aider Configuration

Configure Aider AI coding assistant with Ollama-hosted models and optimized settings for efficient code editing.

What This Skill Does

Sets up Aider with:

Instructions

When configuring Aider for a user:

1. **Create `.aider.conf.yml` in project root** with the following structure:

```yaml

model: ollama_chat/qwen2.5-coder:14b

weak-model: ollama_chat/qwen3:4b

edit-format: udiff

no-show-model-warnings: true

```

2. **Verify Ollama models are available**:

- Ensure `qwen2.5-coder:14b` is pulled: `ollama pull qwen2.5-coder:14b`

- Ensure `qwen3:4b` is pulled: `ollama pull qwen3:4b`

- If models are missing, instruct user to pull them first

3. **Explain configuration choices**:

- **qwen2.5-coder:14b**: Specialized code model, balances capability with speed

- **qwen3:4b**: Lightweight model for simple tasks (commit messages, quick edits)

- **udiff**: Industry-standard unified diff format for code changes

- **no-show-model-warnings**: Reduces noise in output

4. **Note compatibility**: This configuration is for Aider 0.86.1+. Older versions may not support all keys.

5. **Profile management**: If user needs multiple profiles, create separate config files (e.g., `.aider.conf.yml`, `.aider.conf.gpu.yml`) and reference with `--config` flag.

Usage Example

```bash

Launch Aider with this configuration

aider

Or specify config explicitly

aider --config .aider.conf.yml

Override model for specific session

aider --model ollama_chat/codellama:13b

```

Constraints

Notes