Additionally: because this agent is running within a larger workflow, we are looking into tracking the amount of tokens/spend for each workflow run (which has other llm endpoints within the workflow), and part of the spend will be the "estimated cost" from teh agent run.
Are there any plans to release an endpoint for Agents where this data can be accessible?
Another idea that came to mind, could total token usage be added to the agent's response, with the standard breakdown (input tokens, output tokens, reasoning tokens, total tokens)?
With this we can easily take it from the response and store it together with any other token usage from other blocks in the workflow.