Skip to content

📌 Supported Models in PromptRefiner

Promptrefiner leverages LiteLLM under the hood, providing seamless integration with 100+ Large Language Models (LLMs) from top AI providers like OpenAI, Anthropic, TogetherAI, AI21, and more! 🚀

Why does this matter?
This means any model supported by LiteLLM is automatically supported by Promptrefiner. You can refine your prompts using the most advanced AI models available today.

🎯 Supported Models

Since Promptrefiner relies on LiteLLM, it supports all models listed here.

To see the full list of supported models, run:

from litellm import model_list
print(model_list())

🎯 How Model API_KEY Selection Works?

To interact with hosted LLM you will need an API_KEY, it's confidential and mostly provided to any application from environment variable.
When you work with promptrefiner you don't need to worry about it, because we internally map PREFINER_API_KEY to the right API key automatically!
We've designed it to make things easier by allowing a single environment variable:

export PREFINER_API_KEY="your_secret_key_here"

Note

But if you prefer provider-specific API keys, you can still use them! ✅

âš¡ How API Key Handling Works?

Promptrefiner automatically detects the model provider and assigns the appropriate API key.

✔ If using PREFINER_API_KEY → No extra setup needed!
✔ If using provider-specific API keys → The library will detect and use them accordingly.

🛠 Example Usage:

export PREFINER_API_KEY="your_openai_or_anthropic_key"
# If PREFINER_MODEL exported, default to openai/gpt-3.5-turbo
export PREFINER_MODEL="anthropic/claude-3.5"

from promptrefiner import PromptRefiner

prompt_refiner = PromptRefiner(strategies=["persona"])
refined_prompt = prompt_refiner.refine("Explain quantum mechanics.")
print(refined_prompt)

Using Provider-Specific API Keys

export ANTHROPIC_API_KEY="your_anthropic_key"
# If PREFINER_MODEL exported, default to openai/gpt-3.5-turbo
export PREFINER_MODEL="anthropic/claude-3.5"

from promptrefiner import PromptRefiner

prompt_refiner = PromptRefiner(strategies=["persona"])
refined_prompt = prompt_refiner.refine("Explain quantum mechanics.")
print(refined_prompt)