Advanced Aider Configuration

This skill provides a production-ready Aider configuration file (`.aider.conf.yml`) with advanced features including multi-model aliases, OpenRouter integration, customized API settings, and workflow optimizations.

What This Configuration Provides

**Multi-model aliases** for quick switching between providers (GLM, Qwen, Kimi, DeepSeek, Gemini, OpenRouter)

**OpenRouter integration** via custom API base URL

**Optimized model selection** with separate models for main chat, weak tasks, and editing

**Enhanced repo mapping** with 8192 token budget for better code context

**Git workflow customization** with controlled auto-commit behavior

**Dark mode terminal colors** for better readability

**Vim editing mode** for terminal input

**Automatic README inclusion** as read-only context

Instructions

1. **Create the configuration file** in your project's git root or home directory:

```bash

# In project root (recommended for project-specific settings)

touch .aider.conf.yml

# OR in home directory (for global defaults)

touch ~/.aider.conf.yml

```

2. **Copy the configuration** below into `.aider.conf.yml`:

```yaml

alias:

- "glm:openai/GLM-4.5"

- "qc:openai/qwen3-coder"

- "k2:openai/kimi-k2"

- "r1:openai/DeepSeek-R1-0528"

- "v3:openai/DeepSeek-V3-0324"

- "gp:openai/gemini-2.5-pro"

- "gf:openai/gemini-2.5-flash"

- "gfl:openai/gemini-2.5-flash-lite"

- "or:openai/openrouter"

model: gp

weak-model: gfl

chat-mode: ask

openai-api-base: https://liftai.icu/v1

model-settings-file: .aider.model.settings.yml

model-metadata-file: .aider.model.metadata.json

map-tokens: "8192"

dark-mode: true

auto-commits: false

dirty-commits: false

attribute-author: true

attribute-committer: false

read:

- README.md

vim: true

```

3. **Set up your API key** in a `.env` file:

```bash

# Create .env in git root

echo "OPENAI_API_KEY=your_api_key_here" > .env

```

4. **Customize the configuration** for your needs:

- **Change the default model**: Set `model:` to any alias (e.g., `model: r1` for DeepSeek-R1)

- **Adjust repo map tokens**: Increase `map-tokens` for larger codebases (e.g., `"16384"`)

- **Enable auto-commits**: Set `auto-commits: true` if you want Aider to commit changes automatically

- **Add more read-only files**: Add documentation or spec files under `read:`

5. **Create optional model metadata files** (if using custom models):

```bash

# Create placeholder files

touch .aider.model.settings.yml

touch .aider.model.metadata.json

```

6. **Test the configuration**:

```bash

# Launch Aider with the config

aider

# Or switch models using aliases

aider --model r1 # Use DeepSeek-R1

aider --model gf # Use Gemini Flash

```

Key Features Explained

Model Aliases

Quick shortcuts for switching between models without typing full names:

`gp` → Gemini 2.5 Pro (main model)

`gf` → Gemini 2.5 Flash (faster)

`gfl` → Gemini 2.5 Flash Lite (weakest, for summaries)

`r1` → DeepSeek-R1 (reasoning model)

`v3` → DeepSeek-V3 (balanced)

Auto-Commit Behavior

`auto-commits: false` — You manually review and commit changes (recommended)

`dirty-commits: false` — Aider won't commit if repo has uncommitted changes

`attribute-author: true` — Git author reflects Aider's involvement

Repo Mapping

`map-tokens: "8192"` — Allows Aider to understand large codebases with better context

Higher values = more code context, but slower prompt generation

Read-Only Files

Files listed under `read:` are included in every chat as context but won't be edited:

`README.md` — Always available for reference

Add architecture docs, API specs, or style guides here

Usage Examples

**Switch models on the fly:**

```bash

aider --model r1 # Use DeepSeek-R1

aider --model gp --opus # Use Gemini Pro with opus quality

```

**Override map tokens for large repos:**

```bash

aider --map-tokens 16384

```

**Enable auto-commits for prototyping:**

```bash

aider --auto-commits

```

**Add files to chat:**

```bash

aider src/main.py tests/test_main.py

```

Important Notes

**API keys** must be stored in `.env` file (not in YAML for security)

**OpenRouter integration** requires setting `openai-api-base` to your OpenRouter endpoint

**Model metadata files** are optional unless using completely custom models

**Git integration** requires being inside a git repository

**Vim mode** (`vim: true`) enables vi keybindings in the chat interface

Constraints

Only OpenAI and Anthropic API keys can be stored in YAML (use `.env` for others)

Configuration file must be named `.aider.conf.yml` exactly

Must be placed in either git root, current directory, or home directory

Some model aliases require compatible API endpoints

Troubleshooting

**"Model not found" errors:**

Verify your API key in `.env`

Check that `openai-api-base` points to a valid endpoint

Ensure model names in aliases match your provider's model IDs

**Aider not finding config:**

Ensure filename is exactly `.aider.conf.yml`

Check file is in git root (run `git rev-parse --show-toplevel`)

Try explicit path: `aider --config /path/to/.aider.conf.yml`

**Repo map too slow:**

Reduce `map-tokens` to `4096` or `2048`

Use `--map-refresh manual` to disable automatic updates

Advanced Aider Configuration

Advanced Aider Configuration

What This Configuration Provides

Instructions

Key Features Explained

Model Aliases

Auto-Commit Behavior

Repo Mapping

Read-Only Files

Usage Examples

Important Notes

Constraints

Troubleshooting

Reviews (0)