Build a RAG agent with evals using @braintrustdata. Write your functions in Replit, and use the Shell to push to Braintrust. Then, run evals and tweak your prompt and model before you deploy. Full guide in thread.

→ View original post on X — @replit