Skip to main content

​
Installation

pip install respan-exporter-litellm

​
Classes

​
RespanLiteLLMCallback

LiteLLM-native callback handler that sends completion logs to Respan. 
from respan_exporter_litellm import RespanLiteLLMCallback
ParameterTypeDefaultDescription
api_keystr | NoneNoneRespan API key. Falls back to RESPAN_API_KEY env var.
base_urlstr | NoneNoneAPI base URL. Falls back to RESPAN_BASE_URL.

​
Methods

MethodDescription
log_success_event(kwargs, response_obj, start_time, end_time)Log a successful LLM completion.
log_failure_event(kwargs, response_obj, start_time, end_time)Log a failed LLM completion.
async_log_success_event(kwargs, response_obj, start_time, end_time)Async version of success logging.
async_log_failure_event(kwargs, response_obj, start_time, end_time)Async version of failure logging.

​
Helper functions

FunctionDescription
register_litellm_callbacks(name)Register Respan callbacks on the LiteLLM instance.

​
NamedCallbackRegistry

Named registry for managing LiteLLM callbacks.

​
Usage

import litellm
from respan_exporter_litellm import RespanLiteLLMCallback

# Register callback
respan_callback = RespanLiteLLMCallback(api_key="your-api-key")
litellm.callbacks = [respan_callback]

# Use LiteLLM normally — all calls are automatically logged
response = litellm.completion(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Tell me a joke"}],
)
print(response.choices[0].message.content)

​
Async usage

import litellm
from respan_exporter_litellm import RespanLiteLLMCallback

litellm.callbacks = [RespanLiteLLMCallback(api_key="your-api-key")]

response = await litellm.acompletion(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Tell me a joke"}],
)