Walkthrough Guide for Setting Up Agent with Nevron

[gromdimon]

Description

We need a step-by-step walkthrough guide that will help users set up and use Nevron. This guide should be detailed, include commands, screenshots, and videos, and be stored as a Markdown file in the documentation folder.

The guide should cover:

1. Basic setup of Nevron (installation, dependencies).
2. Integration with OpenAI for quick testing. (Add note, that integration of other LLMs is similar)
3. Setting up LLaMA locally and replacing OpenAI with a self-hosted model.
4. Adding integrations (Telegram, Twitter, etc.).
5. Demonstration of successful posts made by the bot (video/screenshots).

This guide should be coherent with index, quickstart and other pages in the documentation.

Requirements

1. Setting Up Nevron

• Install dependencies
• Set up environment variables/configuration (.env).
• Run the agent for the first time (python run.py).

We should chose here only local development without docker

2. OpenAI Integration (Basic Setup)

• Show how to set up an API key for OpenAI.
• Include a basic example of sending queries to the agent.
• Add screenshots of a successful API response.

3. Self-Hosting LLaMA (Replacing OpenAI)

• Run llama in the docker-compose (see the docker-compose.yml for reference)
• Configure Nevron to use a local LLaMA model instead of OpenAI.
• Include screenshots of the self-hosted model running.

4. Adding External Integrations

Telegram Bot Integration

• Set up a Telegram bot using BotFather.
• Connect the bot to Nevron via the Telegram API.
• Demonstrate sending messages through Telegram.

Twitter (X) Integration

• Register a Twitter API key and configure it.
• Post a tweet via the agent.
• Include a screenshot of the bot’s tweet.

5. Real-World Demonstration

• Capture a video or screenshots of the bot posting something successfully.
• Ensure logs confirm the post was made.

---

File Location & Format

• File: somewhere in docs (also add it to the nav in the mkdocs.yml
• Format:
  • Step-by-step commands
  • Screenshots after each major step
  • Code snippets for configurations