Building a philosophy agent
This is a low hanging fruit like project. Crawled SEP to grab entries. Built a UI around it and some standard agent infrastructure behind the scenes. You can see it here on github with instructions on installing and using.
Locally I just use Ollama, but you can also plug in your API key for anthropic. I am not testing it with other LLMs myself. It works pretty well with Ollama.
There is no user based stuff, it is meant to be local. Would be cool to build this into something to make it more widely available but concerned about the Stanford maintainers feeling slighted with all their hardwork. I am going to reach out to them about it. I think making philosophy and the hard work academics have put into the SEP over the years more accessible would be good.
There are two modes, Plato and Socrates. Plato is meant to ask a question and receive answers. Socrates is meant to challenge you and annoy you.
...
You can also hook it up to Claude Desktop, and paste the system prompt for Plato or Socrates.
