Model Selection

Choose the right AI model for your agent based on capabilities, speed, and cost.

Supported Providers

ArcanFlows supports multiple AI providers, giving you flexibility to choose the best model for your use case.

OpenAI

Anthropic

Google

Ollama (Local)

Choosing the Right Model

Decision Framework

                    ┌─────────────────┐
                    │ What's your     │
                    │ priority?       │
                    └────────┬────────┘
           ┌─────────────────┼─────────────────┐
           ▼                 ▼                 ▼
     ┌──────────┐      ┌──────────┐      ┌──────────┐
     │  Speed   │      │ Quality  │      │ Privacy  │
     └────┬─────┘      └────┬─────┘      └────┬─────┘
          ▼                 ▼                 ▼
    GPT-3.5 Turbo     GPT-4 Turbo        Ollama
    Claude Haiku      Claude Opus       (Local)
    Gemini Pro

By Use Case

By Budget

Model Parameters

Temperature

Controls randomness in responses:

Max Tokens

Limits response length:

javascriptCopy
// Short responses (quick answers)
max_tokens: 256

// Medium responses (explanations)
max_tokens: 1024

// Long responses (detailed analysis)
max_tokens: 4096

Top P (Nucleus Sampling)

Controls diversity:

Provider Configuration

OpenAI Setup

1. Get your API key from platform.openai.com
2. Go to Settings > AI Providers
3. Click Add Provider > OpenAI
4. Enter your API key

jsonCopy
{
  "provider": "openai",
  "api_key": "sk-...",
  "organization_id": "org-..." // optional
}

Anthropic Setup

1. Get your API key from console.anthropic.com
2. Go to Settings > AI Providers
3. Click Add Provider > Anthropic
4. Enter your API key

Ollama Setup

For local models:

1. Install Ollama on your server
2. Pull models: ollama pull llama3
3. Configure base URL in ArcanFlows

jsonCopy
{
  "provider": "ollama",
  "base_url": "http://your-server:11434",
  "model": "llama3:70b"
}

Cost Optimization

Tips to Reduce Costs

1. Start with cheaper models

   • Use GPT-3.5 for simple tasks
   • Upgrade only when needed

2. Optimize prompts

   • Shorter prompts = lower costs
   • Remove unnecessary instructions

3. Limit response length

   • Set appropriate max_tokens
   • Ask for concise responses in prompt

4. Cache responses

   • Enable caching for repeated queries
   • Reduces API calls

5. Use local models

   • Ollama for development
   • Ollama for privacy-sensitive data

Cost Comparison (approximate per 1M tokens)

Switching Models

You can switch models at any time:

1. Go to your agent's settings
2. Select Model tab
3. Choose new provider and model
4. Test the agent
5. Publish changes

Note: Different models may respond differently to the same prompt. Always test after switching.

Next Steps

• Knowledge Base - Add context to your agent
• Tools - Enable agent capabilities
• Deployment - Go live with your agent