The primary purpose of Eddie is to demonstrate how to build real-world AI-powered applications. Everything that we use to build Eddie will be open-source and run locally on your machine so it’s private and secure. We will include videos of live coding everything along with detailed written walkthroughs.
You can install Eddie on your own machine by running:
pip install eddie-cli
You can check that Eddie has been installed by running eddie version
.
Before running any command, you’ll need to run ollama run llama3
to spin up your local LLM endpoint for Eddie to query. If you haven’t installed it yet, check out Ollama and install it!
Eddie currently supports the following commands:
eddie version
: outputs the current version of the installed package.eddie hello {name}
: prompt Eddie and say hello.eddie chat
: multi-turn chat with Eddie
[Link to blog post and video]
If there are any particular features or projects you would be interested in watching us build, let us know! In the meantime, this is what’s on our mind roughly:
- Beautiful Textual CLI App (so Eddie looks like an on-board computer)
- RAG with Memory
- Simple Gmail Tool-Use Agent
- Researcher Agent
- Web Search Tools
- Web Scraping Tools
- Technical Documentation Writer
- Code Interpreter
- Interpretable ML Data Scientist (PyTorch Lattice Bot)
- API Documentation Assistant
Eddie uses Semantic Versioning.
This project is licensed under the terms of the MIT License.