Disclaimer Browserman was built during a @modal-labs internal hackathon, and the code has not been cleaned up yet.

Run

• modal deploy llm.py
• modal deploy app.py [or modal serve app.py]
  • Navigate to the URL printed in this step. The LLM takes a couple minutes to cold-start due to the size of the model, but the idle timeout is configured at 20 min so it should stay up for a while.