NexusAI: AI-Powered Strategic Proposal Generator

Welcome to NexusAI, an innovative AI-driven platform built with CrewAI that leverages multi-agent collaboration to generate comprehensive strategic proposals for companies. This project empowers users to analyze market trends, identify AI use cases, collect resources, assess feasibility, and produce detailed proposals—all through an intuitive web interface powered by Streamlit.

Features

• Multi-Agent AI System: Utilizes specialized AI agents for research, use case generation, resource collection, feasibility analysis, and proposal writing.
• Web-Based Interface: Easy-to-use Streamlit app for inputting company details and viewing generated reports.
• Comprehensive Output: Generates detailed Markdown reports and downloadable DOCX files.
• Customizable Agents and Tasks: Easily modify agent roles, goals, and task configurations via YAML files.
• Seamless Integration: Supports various search tools and integrates with external APIs for enhanced data gathering.

Prerequisites

Before setting up the project, ensure you have the following installed on your system:

• Python: Version 3.10 to 3.13 .
• UV: A fast Python package installer and resolver. Install it via:

  pip install uv

• Pandoc: Required for generating DOCX files from Markdown. Download from pandoc.org.
• API Keys: Obtain an OpenAI API key for the AI agents to function properly.

Installation

1. Clone the Repository:

   git clone https://github.com/BlazingPh0enix/NexusAI.git
   cd nexus_ai

2. Install Dependencies: Use UV to install the project dependencies:

   uv sync

   Alternatively, you can use the CrewAI CLI for installation:

   crewai install

3. Set Up Environment Variables: Create a .env file in the project root and add your OpenAI API key:

   OPENAI_API_KEY=your_openai_api_key_here
   

Configuration

Customize the AI agents and tasks to suit your needs:

• Agents Configuration: Edit src/nexus_ai/config/agents.yaml to define agent roles, goals, and backstories.
• Tasks Configuration: Modify src/nexus_ai/config/tasks.yaml to specify task descriptions and expected outputs.
• Core Logic: Update src/nexus_ai/crew.py to add custom logic, tools, or arguments.
• Inputs: Adjust src/nexus_ai/main.py for custom inputs to agents and tasks.

For advanced users, you can integrate additional search tools like Tavily or SerperDev by uncommenting and configuring them in crew.py.

Running the Project

Option 1: Web Interface (Recommended)

Launch the Streamlit web app for an interactive experience:

streamlit run app.py

• Open your browser and navigate to the provided URL (usually http://localhost:8501).
• Enter a company name and industry, then click "Generate Proposal" to create a strategic report.

Option 2: Command Line Interface

Run the CrewAI crew directly from the terminal:

crewai run

This will execute the agents and generate a final_proposal.md file in the output folder with research on Nissan (default example). For custom inputs, modify main.py accordingly.

Usage

1. Input Details: Provide the company name and industry in the web app.
2. Generate Report: Click the "Generate Proposal" button. A spinner will indicate progress.
3. View Results: The report will be displayed in a formatted view with tabs for "Final Report" and "Raw Markdown".
4. Download: Use the download button to save the report as a DOCX file.
5. Customization: Tailor the agents and tasks for specific use cases by editing the configuration files.

Project Structure

nexus_ai/
├── app.py                          # Streamlit web application
├── pyproject.toml                  # Project dependencies and configuration
├── .env                             # Environment variables (not included in repo)
├── uv.lock                         # UV lockfile for dependencies
├── README.md                       # This file
├── output/
│   └── final_proposal.md           # Generated proposal output
├── src/
│   └── nexus_ai/
│       ├── __init__.py
│       ├── main.py                 # Entry point for CLI execution
│       ├── crew.py                 # CrewAI crew definition
│       ├── config/
│       │   ├── agents.yaml         # Agent configurations
│       │   └── tasks.yaml          # Task configurations
│       └── tools/
│           ├── __init__.py
│           └── search_tool.py      # Custom search tools
└── tests/                          # Unit tests

For more information

• 📚 CrewAI Documentation
• 🐙 GitHub Repository