A basic AI Agent framework (WIP) based on Eliza
-
Clone the repo
-
Start poetry shell i.e.
poetry shell
-
Add environment variables - see
.env.example
-
Configure
config.py
to your liking Note this contains logic for the personality, speech, goals and initial thoughts of the agent.
Note code is extremely messy and not well-organised - will clean up at some point.
Some other items to add include
- More sophisticated memory handling
- More sophisticated prompting (esp for the LLM)
- More sophisticated twitter handling (esp for replying to tweets)
- Logging
- Error handling
- Tests
- On chain actions (e.g. using Silverback by ApeWorx : https://github.com/ApeWorX/silverback)