Skip to content

Provider Setup

Amplifier supports multiple LLM providers. This guide covers setting up each one.

Supported Providers

Provider Models Best For
Anthropic Claude 4, Claude 3.5 General purpose, coding, analysis
OpenAI GPT-4, GPT-4o Broad capabilities, function calling
Azure OpenAI GPT-4, GPT-4o Enterprise, compliance requirements
Ollama Llama, Mistral, etc. Local/offline, privacy, experimentation
vLLM Any vLLM-compatible Self-hosted inference, high throughput

For detailed configuration options and advanced features, see Provider Modules.

Get an API Key

  1. Visit console.anthropic.com
  2. Create an account or sign in
  3. Navigate to API Keys
  4. Create a new API key

Configure

# Set environment variable
export ANTHROPIC_API_KEY="sk-ant-..."

# Or add to shell profile
echo 'export ANTHROPIC_API_KEY="sk-ant-..."' >> ~/.bashrc

# Select as provider
amplifier provider use anthropic

Available Models

Model Description
claude-sonnet-4-5 Latest, balanced performance (default)
claude-opus-4-1 Most capable, best for complex tasks
claude-haiku-3-5 Fastest, good for simple tasks 
# Use a specific model
amplifier run --model claude-opus-4-1 "Complex analysis task"

OpenAI

Get an API Key

  1. Visit platform.openai.com
  2. Create an account or sign in
  3. Navigate to API Keys
  4. Create a new secret key

Configure

export OPENAI_API_KEY="sk-..."
amplifier provider use openai

Available Models

Model Description
gpt-4o Latest, multimodal (default)
gpt-4-turbo Fast, capable
gpt-4 Original GPT-4

Azure OpenAI

Prerequisites

  1. Azure subscription
  2. Azure OpenAI resource created
  3. Model deployed in your resource

Configure

export AZURE_OPENAI_API_KEY="your-key"
export AZURE_OPENAI_ENDPOINT="https://your-resource.openai.azure.com"
export AZURE_OPENAI_DEPLOYMENT="your-deployment-name"

amplifier provider use azure

Configuration File (Alternative)

Create ~/.amplifier/settings.yaml:

providers:
  azure:
    api_key: ${AZURE_OPENAI_API_KEY}
    endpoint: https://your-resource.openai.azure.com
    deployment: gpt-4
    api_version: "2024-02-01"

Ollama (Local)

Install Ollama

# macOS
brew install ollama

# Linux
curl -fsSL https://ollama.com/install.sh | sh

# Windows
# Download from https://ollama.com/download

Start Ollama Server

ollama serve

Pull a Model

ollama pull llama3.1
ollama pull codellama
ollama pull mistral

Configure Amplifier

amplifier provider use ollama

Available Models

Any model you've pulled with ollama pull:

# List available models
ollama list

# Use specific model
amplifier run --model llama3.1 "Hello!"

vLLM (Self-Hosted)

For high-throughput self-hosted inference with your own GPU servers.

Prerequisites

  • Running vLLM server (v0.10.1+)
  • Model loaded in vLLM

Configure

# Set your vLLM server URL
export VLLM_BASE_URL="http://your-server:8000/v1"

amplifier provider use vllm

Configuration File

providers:
  vllm:
    base_url: http://192.168.128.5:8000/v1
    default_model: openai/gpt-oss-20b

For vLLM server setup and advanced options, see vLLM Provider.

Switching Providers

Temporary Switch

amplifier run --provider openai "Use OpenAI for this"

Change Default

amplifier provider use anthropic

Check Current Provider

amplifier provider current

Provider Configuration

Via Environment Variables

Provider Variables
Anthropic ANTHROPIC_API_KEY
OpenAI OPENAI_API_KEY
Azure AZURE_OPENAI_API_KEY, AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_DEPLOYMENT
Ollama None required (local)
vLLM VLLM_BASE_URL

Via Settings File

Create ~/.amplifier/settings.yaml:

default_provider: anthropic

providers:
  anthropic:
    api_key: ${ANTHROPIC_API_KEY}
    default_model: claude-sonnet-4-5

  openai:
    api_key: ${OPENAI_API_KEY}
    default_model: gpt-4o

  ollama:
    base_url: http://localhost:11434
    default_model: llama3.1

  vllm:
    base_url: http://your-server:8000/v1
    default_model: openai/gpt-oss-20b

Multiple Providers

You can configure multiple providers and switch between them:

# List configured providers
amplifier provider list

# Switch provider
amplifier provider use openai

# Use specific provider for one command
amplifier run --provider ollama "Local query"

Troubleshooting

"Authentication failed"

  • Verify your API key is correct
  • Check the key hasn't expired
  • Ensure environment variable is exported
echo $ANTHROPIC_API_KEY  # Should show your key

"Model not found"

  • Check the model name is correct
  • For Azure, verify deployment name matches

"Connection refused" (Ollama)

  • Ensure Ollama is running: ollama serve
  • Check it's on the default port: http://localhost:11434

Rate Limits

If you hit rate limits:

  • Wait and retry
  • Use a different model
  • Consider upgrading your API plan

Next Steps