OpenRouter Integration

What it does: Access multiple AI models through a single unified API using OpenRouter. Route requests to various providers like OpenAI, Anthropic, Google, Meta, and more.

When to Use This

Use OpenRouter when you need:

• ✅ Access to multiple AI models without managing separate API keys
• ✅ Cost optimization by choosing the best price/performance ratio
• ✅ Fallback options if one model is unavailable
• ✅ Easy model switching without code changes
• ✅ Access to latest models immediately when released

Example: Build a chatbot that can switch between GPT-4, Claude, Llama, or any other model based on user preference or cost considerations.

Key Features

• Multi-Provider Access: 100+ models from OpenAI, Anthropic, Google, Meta, and more
• Unified API: OpenAI-compatible API format
• Cost Tracking: Built-in usage tracking and billing
• Rate Limiting: Automatic rate limit management
• Model Fallbacks: Switch models if one fails
• Real-Time Updates: Access to newest models as they're released

Setup Guide

Step 1: Get OpenRouter API Key

1. Go to openrouter.ai (opens in a new tab) and sign up
2. Navigate to the API Keys section in your dashboard
3. Click "Create API Key"
4. Give it a descriptive name
5. Copy and save your API key securely

Step 2: Configure the Block

Connection Settings:

1. Credentials: Select your OpenRouter credentials from the dropdown or create new ones

   • API Key: Your OpenRouter API key
   • The system will automatically fetch available models

2. Model: Select the AI model you want to use

   • Dozens of models available (GPT-4, Claude, Llama, etc.)
   • Models are fetched from OpenRouter API
   • Each model has different capabilities and pricing

3. Messages: Configure the conversation

   • System Message: Instructions that guide the AI's behavior
   • User Messages: Add user queries and context
   • Dialogue: Use conversation history from variables

4. Temperature: Control response randomness (0-2)

   • 0: Deterministic, focused responses
   • 1: Balanced creativity (default)
   • 2: Maximum creativity and randomness

5. Response Mapping: Save AI responses to variables

   • Map "Message content" to workflow variables
   • Access total tokens used
   • Store response for later use

Message Configuration

System Message

Define how the AI should behave:

You are a helpful customer support assistant. Be friendly, professional, and concise.
Always ask clarifying questions if the user's request is unclear.

User Messages

Add the user's query:

• Use workflow variables: {{user_question}}
• Combine multiple inputs: User asking about {{topic}}: {{message}}
• Add context: Based on {{previous_answer}}, please explain {{new_topic}}

Dialogue History

Reference conversation history:

• Select a dialogue variable that stores past messages
• Maintains context across multiple exchanges
• Enables natural, coherent conversations

Model Selection

Popular Models Available

OpenAI Models:

• GPT-4 Turbo: Latest OpenAI model
• GPT-4: High capability, reasoning
• GPT-3.5 Turbo: Fast and cost-effective

Anthropic Models:

• Claude 3 Opus: Highest capability
• Claude 3 Sonnet: Balanced performance
• Claude 3 Haiku: Fast and efficient

Open Source Models:

• Llama 3: Meta's latest model
• Mixtral: Mistral AI's mixture of experts
• Command R+: Cohere's command model

Specialized Models:

• Image generation models
• Code-specific models
• Translation models

Common Use Cases

Intelligent Chatbot

Create a responsive assistant that understands context:

• Model: claude-3-sonnet-20240229
• System Message:

  You are a knowledgeable AI assistant. Provide accurate, helpful answers.
  If you don't know something, say so honestly.
• User Message: {{user_question}}
• Temperature: 0.7
• Response Mapping: Save "Message content" to {{ai_response}}

Content Generation

Generate blog posts, product descriptions, or marketing copy:

• Model: gpt-4-turbo
• System Message:

  You are a professional content writer. Create engaging, SEO-friendly content
  that captures the brand voice and appeals to the target audience.
• User Message: Write a blog post about {{topic}}
• Temperature: 1.2 (higher for creativity)

Code Assistant

Help users with programming questions:

• Model: gpt-4-code or claude-3-opus
• System Message:

  You are an expert programmer. Provide clear code examples with explanations.
  Follow best practices and consider security implications.
• User Message: {{code_question}}
• Temperature: 0.3 (lower for accuracy)

Customer Support

Answer customer queries using your knowledge base:

• Model: gpt-3.5-turbo (cost-effective)
• System Message:

  You are a customer support agent. Use the provided context to answer questions.
  Be empathetic, professional, and solution-oriented.
  
  Context: {{knowledge_base_results}}
• User Message: {{customer_question}}
• Temperature: 0.5

Translation Service

Translate content between languages:

• Model: gpt-4 or claude-3-sonnet
• System Message:

  You are a professional translator. Provide accurate translations that preserve
  the meaning, tone, and cultural context of the original text.
• User Message: Translate this from {{source_lang}} to {{target_lang}}: {{text}}
• Temperature: 0.2 (precision matters)

Advanced Configuration

Using Dialogue History

Setup:

1. Create a dialogue variable (e.g., conversation_history)
2. Add messages to it using Set Variable blocks
3. Reference it in OpenRouter block's message section

Example Flow:

1. User sends message → Store in {{user_message}}
2. Append user message to {{conversation_history}}
3. OpenRouter block:
   - Messages: Use dialogue variable {{conversation_history}}
   - Get response
4. Append AI response to {{conversation_history}}
5. Continue conversation with full context

Response Mapping

Extract specific data from AI responses:

Example Mapping:

• Map "Message content" to {{ai_answer}}
• Map "Total tokens" to {{tokens_used}}
• Use {{ai_answer}} in subsequent steps
• Track costs with {{tokens_used}}

What You Get Back

After running an OpenRouter completion:

• Message Content: The AI-generated response
• Total Tokens: Token count for billing/tracking
• Model Used: Confirmation of which model processed the request

Tips for Success

1. Choose the right model - Balance cost, speed, and capability

   • GPT-3.5 Turbo: Fast and cheap for simple tasks
   • GPT-4/Claude 3: Complex reasoning and analysis
   • Llama 3: Open source, good for basic tasks

2. Optimize system messages - Clear instructions improve results

   • Be specific about desired output format
   • Include examples if possible
   • Define tone and style

3. Manage temperature - Adjust based on task

   • Low (0-0.3): Facts, code, structured data
   • Medium (0.5-0.8): Conversations, explanations
   • High (1.0-2.0): Creative writing, brainstorming

4. Use dialogue history - Maintain conversation context

   • Store messages in dialogue variables
   • Include relevant history, not everything
   • Clear context when starting new topics

5. Monitor token usage - Track costs effectively

   • Store token counts in variables
   • Set up alerts for high usage
   • Use cheaper models when possible

Troubleshooting

Best Practices

• Test multiple models - Find the best fit for your use case and budget
• Start with low temperature - Increase only if you need more creativity
• Use system messages effectively - Spend time crafting good instructions
• Implement error handling - Have fallback options if a model fails
• Monitor costs - Track token usage to avoid surprises
• Keep messages concise - Reduce token usage by being clear and brief
• Use streaming - For better user experience in real-time conversations

Pricing

OpenRouter charges based on model used:

• Pay only for what you use
• No subscription fees
• Transparent per-token pricing
• Volume discounts available

Check openrouter.ai/pricing (opens in a new tab) for current rates.