we’ve been building custom LLM-powered agents to perform a wide range of tasks and one of those is screenplay analysis. below are a few screenshots of log output from our agents as they are reading the screenplay text and extracting insights about the story, characters, location, dialogues and more.
here are screenshots from one of our local experiments where we are employing a team of agents (the Krew) to answer a simple question. the agents are built using crewai framework and have access to web search tool. they are also running 100% on local LLMs via Ollama