मुख्य सामग्री पर जाएं

Installation

pip install "rail-score-sdk[litellm]" litellm

Guardrail के रूप में use करें

from rail_score_sdk.integrations import RAILGuardrail
import litellm

# RAIL को guardrail के रूप में register करें
rail_guardrail = RAILGuardrail(
    api_key="YOUR_RAIL_API_KEY",
    threshold=7.0,
    mode="basic",
    action="block",  # "block" | "warn" | "regenerate"
)

litellm.callbacks = [rail_guardrail]

# अब सभी LiteLLM calls automatically score होंगी
response = await litellm.acompletion(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Explain quantum computing."}],
)

# Response metadata से RAIL data access करें
print(response._hidden_params["rail_score"])
print(response._hidden_params["rail_threshold_met"])

Direct usage

from rail_score_sdk.integrations import RAILGuardrail

guardrail = RAILGuardrail(
    api_key="YOUR_RAIL_API_KEY",
    threshold=7.0,
)

result = await guardrail.async_post_call_success_hook(
    data={"messages": [...]},
    response=llm_response,
)
RAILGuardrail LiteLLM का CustomLogger interface implement करता है, तो ये LiteLLM के सभी supported models के साथ काम करता है - Ollama के through local models भी।