Orginal doc on confluence: https://tensormesh.atlassian.net/wiki/spaces/~7120209cca81e6ea95406d80e53f631d9ce9af/pages/745799682/Claude+Code+tracing

Confirmed not working:

claude-trace

LangSmith

claude-code-proxy

https://github.com/badlogic/lemmy

Potential working solution:

MLflow

LiteLLM + Langfuse (looks promising, start with it first)

LiteLLM + Langfuse approach

https://tensormesh.atlassian.net/wiki/spaces/~7120209cca81e6ea95406d80e53f631d9ce9af/pages/745799682/Claude+Code+tracing

langfuse is a piece of *

16:37:47 - LiteLLM:ERROR: litellm_logging.py:4082 - [Non-Blocking Error] Error initializing custom logger: Langfuse.__init__() got an unexpected keyword argument 'sdk_integration' Traceback (most recent call last): File "/Users/kobe/Desktop/lmcache-agent-trace/.venv/lib/python3.12/site-packages/litellm/litellm_core_utils/litellm_logging.py", line 3921, in _init_custom_logger_compatible_class langfuse_logger = LangfusePromptManagement() ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/Users/kobe/Desktop/lmcache-agent-trace/.venv/lib/python3.12/site-packages/litellm/integrations/langfuse/langfuse_prompt_management.py", line 122, in __init__ self.Langfuse = langfuse_client_init( ^^^^^^^^^^^^^^^^^^^^^ File "/Users/kobe/Desktop/lmcache-agent-trace/.venv/lib/python3.12/site-packages/litellm/integrations/langfuse/langfuse_prompt_management.py", line 106, in langfuse_client_init client = Langfuse(**parameters) ^^^^^^^^^^^^^^^^^^^^^^ TypeError: Langfuse.__init__() got an unexpected keyword argument 'sdk_integration'

try local logging

uv pip install litellm uv pip install langfuse

Create config.yaml

`model_list:

• model_name: claude-sonnet-4-5-20250929 litellm_params: model: anthropic/claude-sonnet-4-5-20250929 api_key: os.environ/ANTHROPIC_API_KEY
• model_name: claude-haiku-4-5-20251001 litellm_params: model: anthropic/claude-haiku-4-5-20251001 api_key: os.environ/ANTHROPIC_API_KEY
• model_name: claude-opus-4-5-20251101 litellm_params: model: anthropic/claude-opus-4-5-20251101 api_key: os.environ/ANTHROPIC_API_KEY

litellm_settings: success_callback: ["langfuse"] # logs input/output to Langfuse failure_callback: ["langfuse"] # also log failures`

Set environment variables

`# Your real Anthropic key export ANTHROPIC_API_KEY="sk-ant-***"

LiteLLM master key (any string you choose)

export LITELLM_MASTER_KEY="sk-my-litellm-key"

Langfuse keys (if using Langfuse)

export LANGFUSE_PUBLIC_KEY="pk-lf-" export LANGFUSE_SECRET_KEY="sk-lf-" export LANGFUSE_HOST="https://cloud.langfuse.com"`

Start the proxy