These are the main benefits that you get with requests:

easily tell if the source of the response was from the cache or directly from the LLM
understand which LLM was used
know the tokens count (total, prompt and completion)
clearly see the status code for a request; for example if you get a 400 status code it’s likely that the request was rejected by the LLM due to content filtering or prompt injection attempt
see all the requests associated with a specific cache item

How to view list of all requests

To view list of all requests, navigate to the CogCache CMS and then click on the Requests menu option from the left side. While on this list, you can search them by prompt or filter by status, application or time.

For each request you will have the following information:

status code
date of the request
the prompt
the source, if the response was provided from the LLM or from cache; if answer was from cache it will also display the type of cache hit, if it was an exact or semantic match
the LLM used
the associated cache item; note that if the request is for a first time prompt, the cache item associated is the one that gets created automatically with the request
total tokens
prompt tokens (hidden by default)
completion tokens (hidden by default)

How to export requests

To export requests navigate to the CogCache CMS and then click on the Requests menu option from the left side. Next, go ahead and click on the Export button from top right.

A popup will open that will allow you to export the latest 1,000 requests.

Next, click the Export button. A CSV file with the last 1,000 requests will be downloaded to your computer.