Prevent exposing system prompt with assistant config
I am looking for a way to prevent sending the entire system prompt and other messages to the frontend with the agents. If you could provide an example which does not leak the messages with the agent that would be great.
I have a NextJs fullstack app which needs to take in voice as input -> transcript-> custom llm -> return TTS. I do not want to leak the system prompt/messages to the frontend.
Similar request as this thread; https://discord.com/channels/1211482211119796234/1290311654709399603/1290311654709399603
I have a NextJs fullstack app which needs to take in voice as input -> transcript-> custom llm -> return TTS. I do not want to leak the system prompt/messages to the frontend.
Similar request as this thread; https://discord.com/channels/1211482211119796234/1290311654709399603/1290311654709399603
