Skip to main content

Adapters

Overview​

Adapters in PromptDrifter provide a consistent interface for interacting with different Large Language Model (LLM) providers. They abstract away the provider-specific implementation details, allowing you to seamlessly switch between different LLM providers while maintaining the same code structure.

Supported Adapters​

Adapter Configuration​

Each adapter requires specific configuration parameters. Below are the configuration options for each supported adapter. Note that PromptDrifter supports all models available through each provider's API - the models listed are just examples of what's available.

Common Configuration Options​

These options can be used with any adapter type:

adapter:
  - type: "any-adapter-type"
    model: "model-name"
    skip: true                        # Optional: When true, this adapter will be skipped (default: false)
    temperature: 0.7                  # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                  # Optional: Maximum tokens in the response

OpenAI (GPT)​

OpenAI provides state-of-the-art language models like GPT-3.5 and GPT-4. Their models excel at a wide range of tasks including content generation, summarization, and conversational AI.

Example Models​

The following are examples of available models. PromptDrifter supports any model offered by OpenAI's API:

  • gpt-3.5-turbo - Fast and cost-effective model for most tasks
  • gpt-4 - More capable model with improved reasoning
  • gpt-4o - Latest model with enhanced capabilities across text, vision, and audio
  • gpt-4-turbo - Cost-effective alternative to GPT-4 with similar capabilities

Configuration Options​

adapter:
  - type: "openai"
    model: "gpt-4o"                    # The model to use (default: gpt-3.5-turbo)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 2048                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the OPENAI_API_KEY environment variable.

Claude (Anthropic)​

Anthropic Claude offers a family of AI assistants known for their conversational abilities, reasoning, and instruction-following capabilities.

Example Models​

The following are examples of available models. PromptDrifter supports any model offered by Anthropic's Claude API:

  • claude-3-opus - High-capability model for complex tasks
  • claude-3-sonnet - Balanced model for most tasks
  • claude-3-haiku - Fast, cost-effective model
  • claude-3-5-sonnet - Enhanced model with improved reasoning
  • claude-3-7-sonnet - Latest model with state-of-the-art capabilities

Configuration Options​

adapter:
  - type: "claude"
    model: "claude-3-sonnet"           # The model to use (default: claude-3-opus-20240229)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the CLAUDE_API_KEY environment variable.

Gemini (Google)​

Google's Gemini is a family of multimodal AI models capable of reasoning across text, images, audio, video, and code.

Example Models​

The following are examples of available models. PromptDrifter supports any model offered by Google AI's API:

  • gemini-2.5-pro - Advanced reasoning and multimodal capabilities
  • gemini-2.5-flash - Fast, economical alternative to Pro
  • gemini-2.0-flash-thinking - Specialized for step-by-step reasoning

Configuration Options​

adapter:
  - type: "gemini"
    model: "gemini-2.5-pro"            # The model to use (default: gemini-1.5-flash-latest)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the GEMINI_API_KEY environment variable.

Ollama​

Ollama allows you to run open-source large language models locally. It supports a variety of models including Llama, Mistral, and Gemma.

Example Models​

The following are examples of available models. PromptDrifter supports any model you have installed in your Ollama instance:

  • llama3 - Meta's Llama 3 model
  • mistral - Mistral AI's models
  • gemma - Google's Gemma models
  • Any other models installed in your Ollama instance

Configuration Options​

adapter:
  - type: "ollama"
    model: "llama3"                    # The model to use (default: llama3)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    host: "http://localhost:11434"     # Optional: Ollama host URL
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

Due to Ollama running locally, there is not need to pass in a API key.

Qwen (Alibaba)​

Alibaba Qwen is a series of large language models developed by Alibaba Cloud, known for strong multilingual capabilities and coding performance.

Example Models​

The following are examples of available models. PromptDrifter supports any model offered by Qwen's API:

  • qwen3-30b-a3b - High-capacity model with advanced reasoning
  • qwq-32b - Specialized variant with enhanced instruction following

Configuration Options​

adapter:
  - type: "qwen"
    model: "qwen-plus"                 # The model to use (default: qwen-plus)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the QWEN_API_KEY environment variable.

Grok (xAI)​

xAI Grok is developed and designed to answer questions with humor and personality while having up-to-date knowledge.

Example Models​

The following are examples of available models. PromptDrifter supports any model offered by Grok's API:

  • grok-3 - Latest model with enhanced capabilities
  • grok-2 - Previous generation model

Configuration Options​

adapter:
  - type: "grok"
    model: "grok-1"                    # The model to use (default: grok-1)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the GROK_API_KEY environment variable.

DeepSeek​

DeepSeek offers state-of-the-art language models with strong reasoning and problem-solving capabilities.

Example Models​

The following are examples of available models. PromptDrifter supports any model offered by DeepSeek's API:

  • deepseek-r1 - Research model with advanced reasoning
  • deepseek-v3-0324 - Enterprise-grade model for various applications

Configuration Options​

adapter:
  - type: "deepseek"
    model: "deepseek-chat"             # The model to use (default: deepseek-chat)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the DEEPSEEK_API_KEY environment variable.

Mistral​

Mistral AI provides efficient, high-performance language models with strong reasoning capabilities.

Example Models​

The following are examples of available models. PromptDrifter supports any model offered by Mistral's API:

  • mistral-small-24b-instruct-2501 - 24B parameter model optimized for instruction following
  • mistral-small-3.1-24b-instruct-2503 - Updated 24B parameter model with enhanced capabilities

Configuration Options​

adapter:
  - type: "mistral"
    model: "mistral-large-latest"      # The model to use (default: mistral-large-latest)
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the MISTRAL_API_KEY environment variable.

Azure OpenAI​

Azure OpenAI Service provides enterprise-grade OpenAI models through Microsoft's Azure cloud platform with added security and compliance features.

Example Models​

PromptDrifter supports any model deployed to your Azure OpenAI service:

  • Same models as OpenAI (GPT-3.5, GPT-4, etc.)
  • Custom deployments configured in your Azure account

Configuration Options​

adapter:
  - type: "azure"
    endpoint: "https://your-resource.openai.azure.com/" # Your Azure endpoint
    api_version: "2023-05-15"          # Azure OpenAI API version
    deployment: "your-deployment-name" # Your Azure deployment name
    temperature: 0.7                   # Optional: Controls randomness (0.0-1.0)
    max_tokens: 1024                   # Optional: Maximum tokens in the response
    system_prompt: "Your instruction"  # Optional: System prompt for the conversation

The API key is typically set via the AZURE_OPENAI_API_KEY environment variable.

Adapter Best Practices​

  • API Keys: Store API keys securely using environment variables
  • Error Handling: Implement proper error handling for API rate limits and service outages
  • Caching: Use response caching for frequent identical requests
  • Monitoring: Monitor usage to manage costs and performance