Thread by @CryptoEternalAI

🧠 Decentralized ZerePy Agent 🧠

Merry Xmas from Eternal AI & Zerebro.

This dev guide teaches you how to build your first Decentralized ZerePy agent with:

- Decentralized Inference by @CryptoEternalAI
- Onchain AI Model by @AIatMeta deployed on @Base
- Framework by @0xzerebro

BACKGROUND 1:

For this guide, we'll use @CryptoEternalAI's Decentralized Inference, which has the same interface as OpenAI (so no learning curve for you) — but is decentralized.

We'll use onchain @AIatMeta's Llama 3.3 deployed on @Base.

eternalai.org/api

BACKGROUND 2:

Here is a detailed post if you're curious about how Eternal AI's Decentralized Inference works with @AIatMeta's Llama 3.3 70B and onchain on @Base.

https://x.com/punk3700/status/1869428187450749093

STEP 1

Clone the ZerePy GitHub repo:

github.com/blorm-network/…

STEP 2

Go to the zerepy directory and install dependencies.

> cd zerepy && poetry install --no-root

This will create a virtual environment and install all required dependencies.

STEP 3

Activate the virtual environment.

> poetry shell

STEP 4

Run the application.

> poetry run python

You will see the following output in your terminal. main.py

STEP 5

In the previous step, the ExampleAgentis loaded by default. Let's load the eternalai-example agent.

> agents
> load-agent eternalai-example

STEP 6

Configure Eternal AI connection.

> configure-connection eternalai

STEP 7

Configure Twitter connection.

> configure-connection twitter

Follow the Twitter authentication setup guide to obtain the API Key & Key Secret. Then, enter them into prompts to set up environment variables (stored in a .env file).

STEP 8

Add the TWITTER_USERNAME environment variable to the .env file.

Note: you need to exit the current process before adding it.

> TWITTER_USERNAME="your-twitter-username"

STEP 9

Rerun the agent.

> poetry shell
> poetry run python
> load-agent eternalai-example
> start main.py

STEP 10

With Eternal AI's Decentralized Inference, everything is onchain and verifiable.

Let's look at an "onchain prompt" on @Base Explorer.

This prompt asks the agent to reply to a tweet related to $PENGU by @pudgypenguins.

basescan.org/tx/0x087ae3ece…

STEP 11

Now, let's look at the "onchain response" by the agent.

Everything is onchain and verifiable.

basescan.org/tx/0x3153addd7…

DON'T TRUST (AI). VERIFY.

Here is a detailed post if you're curious about how onchain verification works with Decentralized Inference.

https://x.com/CryptoEternalAI/status/1869611428556218826

If you prefer the developer guide on one single page, here it is:

docs.eternalai.org/eternal-ai/dec…

The future of AI is open-source & decentralized.

The Eternal AI team has been contributing code to the @ai16zdao Eliza framework and is now contributing code to the @0xzerebro ZerePy framework.

If you're working on a cool AI project and need some coding muscle, reach out!