LLM Provider Setup

Set up your AI language model provider to power Fiberwise agents with a single CLI command.

⏱️ 2-5 minutes 📚 Beginner 🔧 Configuration

🤖 Why You Need an LLM Provider

Fiberwise agents use Large Language Models (LLMs) to process and generate intelligent responses. You'll need an API key from one of the supported providers:

🧠 OpenAI

GPT-4, GPT-3.5-turbo, and more

✅ Most popular ✅ Great for general use ✅ Fast responses

🎭 Anthropic

Claude 3, Claude 2, and Haiku models

✅ Excellent reasoning ✅ Safety-focused ✅ Long context

🌊 Cohere

Command and specialized models

✅ Enterprise-ready ✅ Multilingual ✅ Cost-effective

🏠 Local/Self-Hosted

Ollama, vLLM, and custom endpoints

✅ Privacy ✅ No API costs ✅ Full control

⚡ Quick Setup Commands

Choose your provider and run the setup command. The CLI will guide you through the process:

🧠 OpenAI Setup

🧠 Add OpenAI Provider

# Note: LLM provider management is handled through the web UI
# Navigate to http://localhost:5757/settings/llm-providers
# Or manage via the Python SDK in your applications

# Alternatively, providers can be configured via environment variables
export OPENAI_API_KEY="your-api-key-here"

Get your API key: OpenAI API Keys

Models available: gpt-4, gpt-4-turbo, gpt-3.5-turbo

🎭 Anthropic Setup

🎭 Add Anthropic Provider

# Configure Anthropic via web UI or environment variables
export ANTHROPIC_API_KEY="your-api-key-here"

# Or manage through the web interface at:
fiber account set-default-provider anthropic

Get your API key: Anthropic Console

Models available: claude-3-opus, claude-3-sonnet, claude-3-haiku

🌊 Cohere Setup

🌊 Add Cohere Provider

# Add Cohere as your LLM provider
fiber account add-provider cohere --api-key YOUR_COHERE_API_KEY

# Optional: Set as default provider
fiber account set-default-provider cohere

Get your API key: Cohere Dashboard

Models available: command, command-light, command-r

🏠 Local/Self-Hosted Setup

🏠 Add Local Provider

# Add a local or self-hosted provider
fiber account add-provider custom \
  --name "My Local LLM" \
  --endpoint "http://localhost:11434/v1" \
  --api-key "optional-auth-token"

# For Ollama specifically
fiber account add-provider ollama \
  --endpoint "http://localhost:11434" \
  --model "llama2:7b"

Popular local options: Ollama, vLLM, LocalAI, LM Studio

No API key needed for most local setups

✅ Verify Your Setup

Test your provider connection to make sure everything is working:

# List all configured providers
fiber account list-providers

# Test your provider connection
fiber account test-provider openai

# View provider details
fiber account show-provider openai

✅ Expected Output

✅ Providers configured:
  • openai (gpt-4) - Default ✓
  • anthropic (claude-3-sonnet) - Available
  
🧪 Testing openai provider...
✅ Connection successful!
📊 Available models: gpt-4, gpt-4-turbo, gpt-3.5-turbo
💰 Usage: 1,234 tokens used this month

⚙️ Advanced Configuration

🎛️ Provider Settings

Customize your provider settings for optimal performance:

# Set provider-specific defaults
fiber account configure-provider openai \
  --default-model "gpt-4" \
  --max-tokens 2048 \
  --temperature 0.7

# Set rate limits
fiber account configure-provider openai \
  --rate-limit 100 \
  --rate-period "1m"

# Set cost controls
fiber account configure-provider openai \
  --monthly-budget 50.00 \
  --alert-threshold 80

🔐 Environment Variables

You can also set up providers using environment variables:

# Set environment variables
export FIBERWISE_OPENAI_API_KEY="your-api-key"
export FIBERWISE_ANTHROPIC_API_KEY="your-api-key"
export FIBERWISE_DEFAULT_PROVIDER="openai"

# Then configure without exposing keys in command history
fiber account add-provider openai --api-key-from-env

🏢 Team/Organization Setup

For team environments, you can share provider configurations:

# Export provider config for team sharing
fiber account export-providers > team-providers.json

# Import shared provider config
fiber account import-providers team-providers.json

# Or configure organization-wide defaults
fiber org set-default-providers --config org-providers.yaml

🔧 Troubleshooting

❌ "Invalid API Key"

Solution:

Double-check your API key from the provider's dashboard
Ensure no extra spaces or characters
Try regenerating the API key

# Re-add with new key
fiber account update-provider openai --api-key NEW_KEY

❌ "Rate Limit Exceeded"

Solution:

Check your provider's usage dashboard
Upgrade your provider plan if needed
Configure rate limiting in Fiberwise

# Check current usage
fiber account usage-stats openai

❌ "Connection Timeout"

Solution:

Check your internet connection
Verify firewall settings
For local providers, ensure the service is running

# Test with verbose output
fiber account test-provider openai --verbose

❌ "Model Not Found"

Solution:

Check available models for your provider
Ensure you have access to the requested model
Update to a supported model name

# List available models
fiber account list-models openai

🚀 Next Steps

Now that your LLM provider is configured, you're ready to build AI applications!

📚 Start Tutorial

Build your first AI chat app with the Hello World tutorial

Start Building

🤖 Explore Agents

Learn about creating and managing AI agents

Learn Agents

💡 Example Apps

Browse pre-built applications you can deploy

Browse Examples

⚙️ Advanced Config

Deep dive into provider configuration and optimization

Advanced Setup