/bu# Vexel: A Google Cloud Sales Assistant

Vexel is a powerful, AI-driven application designed to assist Google Cloud sales teams. It provides tools for managing customer interactions, analyzing sales data, and streamlining the sales workflow.

This repository contains the backend services and the frontend application for Vexel.

Repository Structure

The project is organized as a monorepo with two main parts:

• / (Backend): The core backend is a Python application built using uv. It contains various agents (calendar_agent, charts_agent, etc.) that handle specific business logic and integrations with Google Cloud services.
• vexel-app/ (Frontend): The frontend is a Next.js application built with pnpm. It provides the user interface for interacting with the Vexel agents and services.

Getting Started

Follow these instructions to set up your local development environment.

Prerequisites

Before you begin, ensure you have the following installed:

• Python: v3.12 or higher
• Node.js: v20 or higher

Environment Setup

1. Install pipx pipx is used to install and run Python applications in isolated environments.

   python3 -m pip install --user pipx
   python3 -m pipx ensurepath

   (You may need to restart your terminal after installing pipx for the changes to take effect.)

2. Install uv uv is an extremely fast Python package installer and resolver.

   pipx install uv

3. Install pnpm pnpm is a fast, disk space-efficient package manager for JavaScript.

   npm install -g pnpm

Backend Setup

1. Create and Activate a Virtual Environment: From the root directory of the project, run:

   python3 -m venv .venv
   source .venv/bin/activate

2. Install Python Dependencies:

   uv pip install -e .

Frontend Setup

Still under development

1. Navigate to the Frontend Directory:

   cd vexel-app

2. Install JavaScript Dependencies:

   pnpm install

3. Run the Development Server:

   pnpm dev

   The application will be available at http://localhost:3000.

ADK Setup

Go back to the root directory for the following.

1. Install the ADK CLI:

   npm i -g @google/generative-ai-adk

2. Login to Google Cloud:

   gcloud auth application-default login

3. Enable the Gemini API:

   gcloud services enable aiplatform.googleapis.com

Configuration

The project uses .env files for managing environment variables.

Backend Configuration (.env)

The Python backend requires credentials for authenticating with Google Cloud services. Create a .env file in the root directory of the project with the following content:

# .env

# Google Cloud Project ID
GOOGLE_CLOUD_PROJECT="your-gcp-project-id"

# Path to your Google Cloud service account key file
# Ensure this file is not committed to version control.
GOOGLE_APPLICATION_CREDENTIALS="/path/to/your/service-account-key.json"

Frontend Configuration (.env.local)

The Next.js frontend requires OAuth client credentials for user authentication. Create a .env.local file in the vexel-app/ directory. You can use the following template:

# vexel-app/.env.local

# Google OAuth 2.0 Client Credentials
OAUTH_CLIENT_ID="your-google-oauth-client-id.apps.googleusercontent.com"
OAUTH_CLIENT_SECRET="your-google-oauth-client-secret"

# The redirect URI must match the one configured in your Google Cloud Console.
REDIRECT_URI="http://localhost:3000/api/auth/callback"

# A secret string for signing session cookies.
SESSION_SECRET="generate-a-strong-random-secret"

graph TD
    A[👤 User] --> B[Pulse Web Application UI];
    B -- HTTPS Request --> C{OrchestratorAgent 🚀 on Cloud Run};
    
    subgraph "Agent Reasoning & Delegation"
        C -- "1. Understands intent" --> D[🧠 Gemini API];
        C -- "2. Delegates task to Sub-Agent" --> E((SessionManagerAgent));
        C -- "2. Delegates task to Sub-Agent" --> F((DataAnalysisAgent));
        C -- "2. Delegates task to Sub-Agent" --> G((RAGAgent));
    end
    
    subgraph "Backend Services & Data"
      E -- "API Call" --> H[Pulse Application API];
      H --> I[(Pulse Production DB)];
      F -- "Semantic Search" --> J[📊 Vertex AI Vector Search];
      J --> I;
      G -- "Retrieves" --> K[RAG Data Sources];
    end

    subgraph "Response Flow"
        H -- "Result" --> C;
        J -- "Result" --> C;
        K -- "Result" --> C;
        C -- "3. Sends Final Response" --> B;
    end

Loading