OpenRouter + Local AI Models Configuration

Multi-provider Aider configuration using OpenRouter API for cloud models (Qwen3 Coder, Kimi K2, Gemini 2.5 Pro, Claude Sonnet 4) and Ollama for local models, with smart model aliasing and dark mode UI.

Instructions

This skill configures Aider with a comprehensive set of model aliases spanning commercial and free OpenRouter models, plus local Ollama instances. It sets Qwen3 Coder (local) as the primary model and Qwen3 as the weak model for faster operations.

Step 1: Verify Prerequisites

Check that Aider, OpenRouter API access, and Ollama (for local models) are available:

```bash

Check Aider installation

aider --version

Verify Ollama is running (if using local models)

ollama list

```

If Ollama models are not installed, pull them:

```bash

ollama pull qwen3-coder:latest

ollama pull qwen3:latest

ollama pull deepseek-coder:1.3b

```

Step 2: Set OpenRouter API Key

Export your OpenRouter API key as an environment variable:

```bash

export OPENROUTER_API_KEY="your_key_here"

```

Add this to your `~/.bashrc`, `~/.zshrc`, or equivalent shell config for persistence.

Step 3: Create Aider Configuration File

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

```yaml

alias:

# Main

- "qwen3-coder:openrouter/qwen/qwen3-coder"

- "kimi:openrouter/moonshotai/kimi-k2"

- "gemini:openrouter/google/gemini-2.5-pro-preview"

- "claude:openrouter/anthropic/claude-sonnet-4"

# Weak

- "qwen3-free:openrouter/qwen/qwen3-coder:free"

- "devstral:openrouter/mistralai/devstral-small-2505:free"

- "chat:openrouter/deepseek/deepseek-r1-0528-qwen3-8b:free"

- "r1:openrouter/deepseek/deepseek-chat-v3-0324:free"

# Local

- "qwen-coder:ollama/qwen3-coder:latest"

- "qwen3:ollama/qwen3:latest"

- "coder:ollama/deepseek-coder:1.3b"

- "orchestrator:ollama/hf.co/jmelahman/Orchestrator-8B-Q4_K_M-GGUF:latest"

model: "qwen-coder"

weak-model: "qwen3"

subtree-only: true

show-model-warnings: false

dark-mode: true

```

Step 4: Test Model Aliases

Launch Aider and verify the default model:

```bash

aider

```

Switch between aliased models using the `/model` command:

```

/model gemini

/model claude

/model qwen3-free

```

Step 5: Understand Model Tiers

**Main models** (high capability, paid):

**Weak models** (fast, free tier):

**Local models** (offline, no API cost):

Step 6: Configuration Options Explained

Usage Examples

**Use default local model:**

```bash

aider

```

**Switch to cloud model mid-session:**

```

/model claude

```

**Force a specific model at launch:**

```bash

aider --model gemini

```

**Use free tier for quick tasks:**

```bash

aider --model qwen3-free --weak-model devstral

```

Notes