How to Use OpenRouter on Mac

OpenRouter is a service that gives you access to 300+ AI models through a single API key. Instead of managing separate accounts with OpenAI, Anthropic, Google, and others, you load credit once and use any model. Here is how to set it up on your Mac.

Step 1: Create an OpenRouter Account

1. Go to openrouter.ai
2. Click “Sign Up” and create an account (Google, GitHub, or email)
3. The process takes about 30 seconds

Step 2: Add Credit

1. Go to your OpenRouter dashboard
2. Navigate to Credits or Billing
3. Add $5-10 to start (you can always add more)
4. Credit is consumed per-token as you use models

How much does $5 get you? Roughly:

• 500+ GPT questions with short answers
• 1,000+ Claude questions
• 5,000+ questions with smaller models like Llama

The exact amount depends on model pricing and response length, but $5 goes a long way for most users.

Step 3: Copy Your API Key

1. In the OpenRouter dashboard, go to Keys
2. Click “Create Key”
3. Name it something like “Chapeta Mac”
4. Copy the key (it starts with sk-or-)

Keep this key private. Anyone with this key can use your credit.

Step 4: Set Up Chapeta

1. Download Chapeta if you have not already
2. Open Chapeta from the menu bar
3. Go to Settings (gear icon)
4. Paste your OpenRouter API key
5. The key is stored in macOS Keychain, encrypted by the OS

That is it. You now have access to every model on OpenRouter.

Step 5: Choose a Model and Start

Click the model selector in Chapeta and browse available models. They are organized by provider:

• OpenAI: GPT, o1, o3
• Anthropic: Claude Opus, Claude Sonnet, Claude Haiku
• Google: Gemini Pro, Gemini Flash
• Meta: Llama (various sizes)
• Mistral: Mistral Large, Codestral
• DeepSeek: DeepSeek R1, DeepSeek V3

Pick a model and type your first message. The response comes back in the same interface regardless of which model you chose.

Model Selection Tips

• Everyday questions: Claude Haiku, Gemini Flash, or Llama - fast and cheap
• Complex reasoning: Claude Opus, GPT, or o1 for deep analysis
• Coding: DeepSeek or Claude Sonnet
• Long documents: Claude (200K context) or Gemini Pro (1M context)
• Cost efficiency: Llama and other open-source models cost a fraction of premium models

You can switch models between messages. Start a conversation with a cheap model and switch to a premium one when you need deeper analysis.

Managing Costs

OpenRouter shows your usage and remaining credit in the dashboard. Tips for managing costs:

• Set a budget alert: OpenRouter lets you set spending limits
• Use small models for small tasks: Do not use a premium model for a simple question that a lightweight model can handle
• Watch token counts: Longer conversations cost more. Start fresh conversations for new topics
• Check model pricing: OpenRouter lists per-token pricing for every model. Premium models like Claude cost significantly more than models like Llama

Troubleshooting

“Invalid API key”: Make sure you copied the full key including the sk-or- prefix.

“Insufficient credit”: Add more credit to your OpenRouter account.

“Model unavailable”: Some models have capacity limits. Try again in a moment or switch to an alternative model.

Slow responses: Response speed depends on the model and current load. Smaller models respond faster. If a model is consistently slow, try an alternative from the same provider.