Chat with a local AI model via Ollama
The sandboxed piston, for code executions
Parsing the document with super high accuracy