Appearance

Environment Setup ​

When you clone the repo there is a sample environment setup in sample.env. You can configure variables for the models used in your simulation, your database, and to control the level of concurrency.

Required Model Configuration ​

  • API keys
  • LLM_PROVIDER sets the api (options are "openai", "gemini", "anthropic")
  • LLM_MODEL sets the model (e.g. "gpt-4.1")

Optional Extra Model Configuration ​

  • LLM_REASONING_EFFORT controls the depth of reasoning for models that support it (options: "minimal", "standard", "high")
  • LLM_TEMPERATURE controls response randomness
  • LLM_MAX_TOKENS sets maximum tokens generated per response
  • LLM_MAX_CONCURRENCY limits concurrent requests happening at once to prevent rate limiting

Database Setup ​

There are several variables in the sample.env that must be set to determine database login. We encourage you to use the defaults. In addition you can set:

  • POSTGRES_MAX_CONNECTIONS limits the number of simultaneous connections in the pool

FAQ ​

  • How can I prevent rate limiting errors?

    Try reducing the LLM_MAX_CONCURRENCY to something like 10 (setting to 1 means that each LLM call will happen sequentially).

  • How can I fix errors related to too many database connections?

    Try reducing your POSTGRES_MAX_CONNECTIONS.

  • How can I run more simulations in parallel with the same database?

    Try reducing your POSTGRES_MAX_CONNECTIONS.