-
Hi, Thank you for this amazing project! I was wondering if it's possible to export the history of LLM calls (including both inputs and outputs) as a file during the process. This feature would be highly useful for tracking and debugging model interactions. If this functionality isn’t available, are there any recommended approaches or tools to implement such logging? Thanks in advance for your guidance! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
Welcome @naufalso, A) Langsmith can help: You'll get an API key there. If you set these 2 env variables, you should get logs on cost and errors within your Langsmith Dashboard (calculated per LLM call or grouped by project) LANGCHAIN_TRACING_V2=true B) regarding the GPTR flow logs: https://docs.gptr.dev/docs/gpt-researcher/gptr/handling-logs-as-they-stream |
Beta Was this translation helpful? Give feedback.
Welcome @naufalso,
A) Langsmith can help:
https://smith.langchain.com/
You'll get an API key there.
If you set these 2 env variables, you should get logs on cost and errors within your Langsmith Dashboard (calculated per LLM call or grouped by project)
LANGCHAIN_TRACING_V2=true
LANGCHAIN_API_KEY=Set this to your API key
B) regarding the GPTR flow logs:
https://docs.gptr.dev/docs/gpt-researcher/gptr/handling-logs-as-they-stream