tstAgent

Hi, I hope you are doing great, guys. I am currently working on this project. I am sure those who are building agents are having a hard time testing their agents because of hallucinations and bad prompting. We have this proposal: if you tell us your agent’s goal, we can create test scenarios and test your agent. This way, you are able to know which cases were successful and which cases weren’t, so you can fix it. Once your agent is deployed, you can still monitor the agent’s performance and see if it’s completing the given goal or not. We are in beta testing. If you are interested in joining us, we are happy to collaborate.