🔍 BrandOS

Realtime Brand Monitoring and Analysis

Overview • Features • Prerequisites • Installation • Usage • Functions • License

BrandOS is a Python script that utilizes the Anthropic API and You Web LLM Search to monitor, analyze, and suggest actions for a brand to maintain unique differentiation. This project is based on Maestro by Doriandarko.

📖 Overview

BrandOS contains the code and instructions needed to build a sophisticated brand monitoring and analysis tool that leverages the capabilities of Anthropic API and You Web LLM Search. Designed to efficiently break down objectives into sub-tasks, generate prompts for sub-agents, and refine the results into a cohesive final output, this project is an ideal starting point for developers interested in brand management and natural language processing.

🎛️ Features

Utilizes the Anthropic API for natural language processing and generation
Integrates with the You Web LLM Search API for QnA search functionality
Breaks down the objective into sub-tasks using an orchestrator
Generates prompts for sub-agents to execute the sub-tasks
Refines the sub-task results into a cohesive final output
Provides a list of competitors, keyword search terms to monitor, and suggested actions for the brand
Saves the full exchange log to a Markdown file

📋 Prerequisites

Before running the script, make sure you have the following:

Python 3.x installed
Required Python packages: os, re, json, dotenv, anthropic, rich, datetime
Anthropic API key (stored in a .env file as ANTHROPIC_API_KEY)
You Web LLM Search API key (stored in a .env file as You Web LLM Search_API_KEY)

🚀 Installation

Clone the repository or download the script file.

Install the required Python packages by running the following command:

pip install os re json dotenv anthropic You Web LLM Search rich datetime

Create a .env file in the same directory as the script and add your Anthropic and You Web LLM Search API keys:
```
ANTHROPIC_API_KEY=your_anthropic_api_key
You Web LLM Search_API_KEY=your_You Web LLM Search_api_key
```

💻 Usage

Run the script using the following command:
```
python brandos.py
```
When prompted, enter the following information about the brand:
- Brand URL
- Brand mission
- Brand vision
- Brand values
Choose whether you want to use search functionality by entering 'y' or 'n'.
The script will start the process of monitoring, analyzing, and suggesting actions for the brand. It will break down the objective into sub-tasks, generate prompts for sub-agents, and refine the results.
The final output will be displayed in the console, including a list of competitors, keyword search terms to monitor, and suggested actions for the brand.
The full exchange log will be saved to a Markdown file with a timestamp and the sanitized brand name.

🧩 Functions

calculate_subagent_cost(model, input_tokens, output_tokens): Calculates the cost of a sub-agent based on the model and token usage.
opus_orchestrator(brand_info, previous_results=None, use_search=False): Calls the orchestrator to break down the objective into sub-tasks and generate prompts for sub-agents.
haiku_sub_agent(prompt, search_query=None, previous_haiku_tasks=None, continuation=False): Calls the sub-agent to execute a sub-task based on the provided prompt and search query.
opus_refine(brand_info, sub_task_results, filename, continuation=False): Calls Opus to review and refine the sub-task results into a cohesive final output.

📄 License

BrandOS is open-source and licensed under MIT.

amadad/brandOS