What feature would you like to be added?
- User should be able to understand LLM cost of an end to end run
- User should be able to track cost within a context (requires propagation between publishes. Look into how tracing is achieving this)
- User should be able to cancel a run early based on cost
Why is this needed?
API cost can be expensive and keep track of usage is important for production