ZerePy is an open-source Python framework designed to let you deploy your own AI agents on X (formerly Twitter), powered by OpenAI or Anthropic LLMs. The framework also supports a ton of blockchain actions, some of which are unique to ZerePy in the agent space.
The GitHub repo has over 500 stars and has been used as a base in multiple hackathons.
My main contributions to this project were designing the initial class structure for ZerePy, adding basic Twitter and OpenAI/Anthropic support, and designing the LangGraph decision flow which allows agents to orchestrate sequential actions from plain text requests, make decisions based on previous actions and results, and even operate autonomously without human input.
This project was created while working with the Blorm team.
Links:
Motivations:
I was excited to work on this project because I did not have any open-source project experience and I wanted to explore what was possible with LLM-powered agents.
Implementation:
This framework is built with Python.
Future Work:
Future work can be found on the repo itself in the form of issues and feature requests.