Message structure: list[<role>: <content>].

Wrapper around SerpAPI. To use, you should have the ``google-search-results`` python package installed, and the environment variable ``SERPAP

OpenAI-compatible assistant message.

Typical usage: config = Config("config.yaml") secret_key = config.get_key("MY_SECRET_KEY") print("Secret key:", secret_key)

Track API usage cost.

The memory storage with Faiss as ANN search engine

Raised when the operation cannot be completed due to insufficient funds

Runtime context for a role.

Role configuration settings.

TODO: Integrate Google Search and handle proxying if needed. Note: Google may require a system/global proxy setup. - DDG: https://pypi.or

Wrapper around SerpAPI. To use, you should have the ``google-search-results`` python package installed, and the environment variable ``SERPAP

OpenAI-compatible system message.

OpenAI-compatible user message.

Architect: Listen to PRD, responsible for designing API, designing code files

Abstract GPT class

GPT API abstract class, requiring all inheritors to provide a series of standard capabilities

FIXME: Consider add_index, set_index, and appropriate granularity.

Environment hosting multiple roles; roles publish messages here, observable by others.

Unified LLM provider using LiteLLM for routing.

Long-term memory component for a role. - Restore historical memory for the role at startup - Update storage when memory changes

The most basic memory: super-memory

Exception raised for errors in the configuration. Attributes: message -- explanation of the error

Simple RPM-based rate limiter.

Requirement without any implementation details

Role/Agent.

Singleton metaclass for ensuring only one instance of a class.