# Project export: MicroMentor This document was generated by HackStack to give an AI agent context about a hackathon project. Sections are labeled with their provenance; content marked as truncated was cut to keep this document small. ## Project metadata - Hackathon: OpenAI Build Week - Tagline: Saving micro-businesses from failing through bilingual, agentic AI workflows. - Devpost: https://devpost.com/software/micromentor-souf06 - GitHub: https://github.com/cryptXploit/MicroBusinessMentor.ai.git - Video: https://www.youtube.com/embed/4LW7L6CG180?enablejsapi=1&hl=en_US&rel=0&start=&version=3&wmode=transparent - Team: 1 GitHub contributor(s) — cryptXploit (6 commits) ## Devpost submission (written by the team) ### Overview 💡 ### Inspiration In regions like Rajshahi, Bangladesh, micro-entrepreneurs—ranging from seasonal mango orchard owners to local boutique operators—face an alarming 80% failure rate within their first year. They don't fail due to lack of grit; they fail due to a lack of strategic foresight, financial literacy, and immediate crisis management. While enterprise corporations have access to seasoned CFOs, legal teams, and market analysts, micro-business owners operate in total isolation. MicroMentor was born out of a single question: What if a local street vendor or small agro-business could afford a Fortune 500-level Board of Directors for free? Our mission is to democratize high-level business intelligence by combining Agentic AI workflows, native bilingual support (English & Bengali), and accessible low-latency architectures. 🏗️ How We Built It MicroMentor was engineered from the ground up using a Hub-and-Spoke architecture on Next.js 14, paired with Supabase for secure data isolation and real-time state management. ⚡ Powered by OpenAI Codex & GPT-5.6 Codex as the Execution Engine: We adopted a Pre-planned Prompt Architecture methodology. Instead of typing vague queries, we drafted structured system specifications locally and passed them to OpenAI Codex. Codex handled ~80% of our frontend Glassmorphism UI components, Tailwind styling, and Next.js API routing setup, drastically accelerating development. Dynamic Model Routing ($GPT\text{-}5.6$): To deliver deep reasoning without frustrating latency, we implemented a dynamic router between the GPT-5.6 model family based on task complexity: $GPT\text{-}5.6\text{-}Luna$ (Deep Reasoning): Powers our multi-agent War Room (AI Board of Directors), SOS Crisis Manager, and Legal Desk where zero-shot accuracy and complex trade-off evaluation are required. $GPT\text{-}5.6\text{-}Mini$ (Flash Execution): Powers low-latency daily utilities such as AI Khata (Voice-to-Ledger), Smart SMS, and Customer Roleplay. 🧮 Model Routing & Latency Optimization Formula To optimize the balance between execution speed and intelligence quality, we defined the routing decision utility function $U(M)$ for selecting model $M \in {\text{Luna}, \text{Mini}}$ given a user request $R$: $$U(M) = w_1 \cdot \text{ReasoningDepth}(R, M) - w_2 \cdot \text{Latency}(M) - w_3 \cdot \text{Cost}(M)$$ Where weights $w_1, w_2, w_3$ dynamically shift depending on whether the feature is real-time interactive (e.g., chat) or deep analytical (e.g., legal drafting). 🚧 Challenges We Faced Latency vs. User Perception: Deep multi-agent reasoning in the AI Board of Directors takes processing time. To prevent user drop-off, we built an Agentic Streaming UI that displays real-time thought logs ("Analyzing supply chain...", "Consulting Financial Advisor...") to mask execution latency smoothly. Bilingual Context Precision: Translating complex business logic into localized Bengali without losing technical context (e.g., profit margin calculations or partnership clauses) required fine-tuning system prompts and structured JSON outputs. Push Protection & Security: Managing secrets while maintaining rapid git cycles during the hackathon forced us to implement strict pre-commit verification workflows. 📚 What We Learned AI as a Pair Programmer: Using OpenAI Codex with strict architectural constraints yields production-grade code in minutes rather than hours, provided the developer designs the system boundaries first. Designing for Micro-Users: True accessibility isn't just about dark mode or glassmorphism—it’s about enabling voice inputs, native regional language understanding, and actionable, one-click execution plans. 🚀 What's Next for MicroMentor WhatsApp & SMS Bot Integration: Allowing vendors to update their ledger and query their AI Co-Founder via basic feature phones without requiring continuous internet access. Localized Logistics Integration: Connecting local transport systems directly with the Autonomous Lead Sniper to automate supply chain fulfillment. ## README (from the GitHub repository) # 🚀 MicroMentor: The Enterprise-Grade AI Co-Founder for Micro-Entrepreneurs [![Next.js](https://img.shields.io/badge/Next.js-14-black?style=for-the-badge&logo=next.js)](https://nextjs.org/) [![OpenAI](https://img.shields.io/badge/OpenAI-GPT--5.6%20%7C%20Sora%20%7C%20DALL--E%203-412991?style=for-the-badge&logo=openai)](https://openai.com/) [![Supabase](https://img.shields.io/badge/Supabase-Database-3ECF8E?style=for-the-badge&logo=supabase)](https://supabase.com/) [![Tailwind CSS](https://img.shields.io/badge/Tailwind-CSS-38B2AC?style=for-the-badge&logo=tailwind-css)](https://tailwindcss.com/) **MicroMentor** is an advanced, multi-modal, agentic AI platform designed to democratize enterprise-level business intelligence for micro-entrepreneurs. It acts as a 24/7 AI Co-Founder, Business Strategist, and Legal/Financial Advisor. --- ## ⚠️ The Problem: The "Death Valley" of Micro-Businesses Across the globe, particularly in developing economies like Bangladesh, **80% of micro-businesses shut down within their first year**. Why? 1. **Financial Illiteracy:** Founders don't understand cash flow, burn rate, or ledgers. 2. **Lack of Strategy:** Decisions are made on gut feelings, not data or market trends. 3. **Resource Scarcity:** Hiring a marketing agency, a corporate lawyer, or a business consultant is financially impossible. 4. **Language & Tech Barriers:** Existing enterprise tools (like Salesforce or SAP) are too complex, expensive, and English-only. ## 💡 The Solution: MicroMentor MicroMentor bridges this massive gap by providing a **Bilingual (English & Bengali), Mobile-first, Hub-and-Spoke Ecosystem**. We utilize **GPT-5.6, Agentic Workflows, RAG (Live Web Search), and Multi-Modal AI (Sora/DALL-E 3)** to give a local mango seller in Rajshahi the same strategic power as a Fortune 500 CEO. --- ## 🌟 The 12 "God-Tier" Features (What makes us stand out) ### 🧠 Agentic Intelligence & Strategy * **⚖️ Multi-Agent War Room:** When facing a crisis, summon an AI Board of Directors. 3 distinct personas (The Conservative CFO, The Aggressive CMO, and The Wise CEO) debate in real-time to provide a balanced verdict. * **🎯 Autonomous Lead Sniper:** Type your product, and the AI autonomously scours the internet, finds 3-4 real-world B2B companies/buyers, and drafts hyper-personalized outreach pitches. * **🕵️ Live Competitor Spy:** Uses real-time DuckDuckGo web scraping to analyze current market prices, find local competitors, and cite verified sources. ### 💰 Finance & Compliance * **⏳ Financial Time Machine (Burn-Rate Simulator):** An interactive slider that calculates runway. If bankruptcy is imminent (<3 months), it triggers an **AI Survival Pivot** to save the business. * **📓 AI Khata (Voice-to-Ledger):** Users can simply speak or type messy daily transactions (e.g., "Sold 2 shirts for $20, bought packaging for $5"). The AI structures it into a professional Balance Sheet. * **⚖️ AI Legal Desk:** Drafts court-ready legal documents (Partnership Deeds, NDAs, Trade License forms) customized to the specific business, ready to print. ### 🚀 Marketing & Media * **🎬 Viral Marketing Studio:** Generates cinematic prompt engineering (camera angles, lighting, psychological hooks) for 30-sec viral reels and graphic posts. * **🪄 End-to-End Media Generator (FAB):** A smart Floating Action Button integrated with **DALL-E 3** and **OpenAI Sora**. It maintains "Consistent Brand Characters" across sequential video generations directly within the app. * **💬 Smart SMS Communicator:** Drafts professional B2B/B2C texts for difficult scenarios (e.g., angry customer apologies, supplier negotiations). ### 🔮 Data & Ecosystem * **🔮 Predictive Radar:** Generates a 12-month local trend and seasonality heatmap based on geographic location and weather patterns, providing pre-emptive tasks. * **🚚 Supply Chain Matchmaker:** Finds real-world wholesale markets and suppliers using live web search, providing negotiation tips. * **🔊 The Mentor's Voice & Bilingual Core:** OpenAI TTS integration with a 100% lightweight bilingual architecture (`?lang=bn`) for deep local penetration. --- ## Section Title: 🛠️ Development Workflow: How I Used Codex & GPT-5.6 As a developer with a background in Full-Stack Engineering and Data Structures, I knew that relying on AI blindly leads to spaghetti code and excessive DOM bloat. Therefore, I treated OpenAI Codex not just as a chatbot, but as an execution engine. Instead of writing vague queries in the Codex prompt box, I adopted a strict 'Pre-planned Prompt Architecture' methodology. Here is my exact workflow: System Design First: Before writing a single line of code, I mapped out the entire Next.js 14 App Router structure, the Hub-and-Spoke architecture, and the Supabase database schema offline. Offline Prompt Drafting: To ensure zero-shot success and avoid wasting context windows, I drafted highly structured, detailed prompts in my local code editor first. I explicitly defined the UI requirements (Glassmorphism, Tailwind), routing logic, and state management rules in these text files. Codex Execution: I then pasted these hyper-structured prompts into Codex. This is why my Codex session logs show large, well-formatted prompt pastes. Codex acted as my senior pair-programmer—it took my architectural instructions and rapidly generated the modular components, drastically cutting down my frontend UI development time by 80%. GPT-5.6 for the 'Brain': While Codex built the body (the app), I integrated GPT-5.6-Luna and GPT-5.6-Mini via API to act as the 'brain'. I used Luna for heavy, multi-agent reasoning (like the War Room and SOS Crisis) and Mini for low-latency tasks (like AI Khata and Smart SMS). Human-in-the-Loop Refinement: After Codex generated the initial codebase, I manually refined the routing middleware, fixed Next.js hydration issues, and polished the bilingual (English/Bengali) translation logic to ensure a seamless UX. --- ## 🏗️ System Architecture We implemented a **"Hub and Spoke" Architecture** to prevent DOM Bloat and ensure extreme performance on low-end devices: * **The Hub:** A sleek, gamified Dashboard (`/dashboard`) acting as the AI Command Center. * **The Spokes:** Isolated Next.js routes (`/radar`, `/war-room`, `/legal-desk`) for heavy AI tools, ensuring zero latency on the main dashboard. * **Agentic Streaming UI:** Masks API latency with dynamic "thought-process" loading states (e.g., *Deploying AI Agent... -> Fetching local data...*). --- ## 💻 Tech Stack * **Frontend:** Next.js 14 (App Router), React, Tailwind CSS, Lucide Icons, Canvas-Confetti (Magic moves). * **Backend:** Next.js Route Handlers (API). * **AI & ML:** OpenAI SDK (`gpt-5.6`, `gpt-5.6-mini`, `dall-e-3`, `sora` concept), Web Speech API. * **Search/RAG:** DuckDuckGo Web Scraper Utility. * **Database:** Supabase (PostgreSQL) with JSONB structures. --- ## 🚀 Installation & Local Setup ### Prerequisites * Node.js (v18.x or later) * An OpenAI API Key (with GPT-5.6 access) * A Supabase Project ### 1. Clone the Repository ```bash git clone [https://github.com/your-username/micro-mentor.git](https://github.com/your-username/micro-mentor.git) cd micro-mentor 2. Install Dependencies Bash npm install # or yarn install 3. Setup Environment Variables Create a .env.local file in the root directory and add the following keys: Code snippet NEXT_PUBLIC_SUPABASE_URL=your_supabase_project_url NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_key OPENAI_API_KEY=your_openai_api_key 4. Database Setup (Supabase SQL Editor) Run the following SQL commands in your Supabase SQL Editor to set up the necessary tables and JSONB columns: ```sql -- WARNING: This schema is for context only and is not meant to be run. -- Table order and constraints may not be valid for execution. CREATE TABLE public.users ( id uuid NOT NULL, email text NOT NULL UNIQUE, created_at timestamp with time zone NOT NULL DEFAULT timezone('u [README truncated for size] ## Detected evidence (automated analysis) Indexed codebase: 72 recognized source files, 522 KB. - CSS (language) — detected in the code - Next.js (technology) — detected in the code - OpenAI (technology) — detected in the code - React (technology) — detected in the code - SQL (language) — detected in the code - Supabase (technology) — detected in the code - Tailwind CSS (technology) — detected in the code - TypeScript (language) — detected in the code - JavaScript (language) — claimed on Devpost, not found in the code - Node.js (technology) — claimed on Devpost, not found in the code - PostgreSQL (technology) — claimed on Devpost, not found in the code ## Codebase structure (from repository index) ### Files (81 of 81) ``` .env.example .eslintrc.json .gitignore components.json next.config.mjs package.json postcss.config.mjs README.md src/app/ai-khata/page.tsx src/app/api/deep-dive/route.ts src/app/api/financial-pivot/route.ts src/app/api/generate-audio/route.ts src/app/api/generate-media/route.ts src/app/api/generate-pitch/route.ts src/app/api/generate-script/route.ts src/app/api/generate/route.ts src/app/api/khata-transcribe/route.ts src/app/api/khata/route.ts src/app/api/lead-sniper/route.ts src/app/api/legal/route.ts src/app/api/marketing-brand/route.ts src/app/api/marketing-character/route.ts src/app/api/marketing-content/route.ts src/app/api/marketing-studio/route.ts src/app/api/phase-two/route.ts src/app/api/questions/route.ts src/app/api/radar-search/route.ts src/app/api/radar/route.ts src/app/api/roleplay/route.ts src/app/api/smart-sms/route.ts src/app/api/sos/route.ts src/app/api/speak/route.ts src/app/api/supply-chain/route.ts src/app/api/survival-pivot/route.ts src/app/api/task-solution/route.ts src/app/api/war-room/route.ts src/app/consultation/page.tsx src/app/dashboard/page.tsx src/app/globals.css src/app/layout.tsx src/app/lead-sniper/page.tsx src/app/legal-desk/page.tsx src/app/login/page.tsx src/app/page.tsx src/app/radar/page.tsx src/app/services/mockAiService.ts src/app/signup/page.tsx src/app/smart-sms/page.tsx src/app/supply-chain/page.tsx src/components/ActionPlan.tsx src/components/AgenticLoader.tsx src/components/AgentReasoningLoading.tsx src/components/AudioGenerator.tsx src/components/AuthPanel.tsx src/components/BudgetChart.tsx src/components/BusinessInsights.tsx src/components/DashboardSkeleton.tsx src/components/FinancialTimeMachine.tsx src/components/FloatingAssetGenerator.tsx src/components/FundingPitch.tsx src/components/MarketingStudio.tsx src/components/Navbar.tsx src/components/RiskAlerts.tsx src/components/RoleplaySimulator.tsx src/components/ui/button.tsx src/components/ui/card.tsx src/components/ui/checkbox.tsx src/components/ui/input.tsx src/components/ui/progress.tsx src/components/WarRoom.tsx src/data/case_studies.json src/lib/utils.ts src/types/index.ts src/utils/mockBusinessPlan.ts src/utils/supabase/client.ts src/utils/translations.ts supabase.sql supabase/schema.sql supabase/semantic-cache.sql tailwind.config.ts tsconfig.json ``` ### Dependencies - package.json: @base-ui/react@^1.6.0, @supabase/supabase-js@^2.110.4, @types/canvas-confetti@^1.9.0, @types/node@^20, @types/react@^18, @types/react-dom@^18, canvas-confetti@^1.9.4, class-variance-authority@^0.7.1, clsx@^2.1.1, eslint@^8, eslint-config-next@14.2.35, lucide-react@^1.24.0, next@14.2.35, openai@^6.46.0, postcss@^8, react@^18, react-dom@^18, recharts@^3.9.2, shadcn@^4.13.0, tailwind-merge@^3.6.0, tailwindcss@^3.4.1, tailwindcss-animate@^1.0.7, tw-animate-css@^1.4.0, typescript@^5 ### Recent commits (newest first) - Add example environment variables to .env.example - Document development workflow with Codex and GPT-5.6 - Revise README for MicroMentor project - Removed Chatgpt Some prompt what i generated for Codex resources folder - Final Release: MicroMentor Hackathon 2026 - Initial commit from Create Next App ## Key source files (fetched from GitHub, selected and truncated for size) ### package.json ``` { "name": "micro-business-mentor", "version": "0.1.0", "private": true, "scripts": { "dev": "next dev", "build": "next build", "start": "next start", "lint": "next lint" }, "dependencies": { "@base-ui/react": "^1.6.0", "@supabase/supabase-js": "^2.110.4", "canvas-confetti": "^1.9.4", "class-variance-authority": "^0.7.1", "clsx": "^2.1.1", "lucide-react": "^1.24.0", "next": "14.2.35", "openai": "^6.46.0", "react": "^18", "react-dom": "^18", "recharts": "^3.9.2", "shadcn": "^4.13.0", "tailwind-merge": "^3.6.0", "tailwindcss-animate": "^1.0.7", "tw-animate-css": "^1.4.0" }, "devDependencies": { "@types/canvas-confetti": "^1.9.0", "@types/node": "^20", "@types/react": "^18", "@types/react-dom": "^18", "eslint": "^8", "eslint-config-next": "14.2.35", "postcss": "^8", "tailwindcss": "^3.4.1", "typescript": "^5" } } ``` ### src/app/layout.tsx ```typescript // app/layout.tsx import type { Metadata } from "next"; import { Inter } from "next/font/google"; // @ts-ignore import "./globals.css"; import Navbar from "@/components/Navbar"; const inter = Inter({ subsets: ["latin"] }); export const metadata: Metadata = { title: "MicroMentor AI", description: "Your AI-powered micro-business mentor.", }; export default function RootLayout({ children, }: Readonly<{ children: React.ReactNode; }>) { return ( {children} ); } ``` ### src/types/index.ts ```typescript export interface BusinessPlan { businessName: string; totalBudget: number; allocatedBudget: Record; } export interface ActionStep { timeframe: string; task: string; isCompleted: boolean; mentorSolution?: string; generatedScript?: string; } export interface RiskAlert { title: string; description: string; } export interface AgentStep { id: string; label: string; detail: string; status: "pending" | "active" | "complete"; } export interface UnitEconomics { product: string; costToMake: number; sellingPrice: number; profitPerUnit: number; breakEvenUnits: number; microBOM: { item: string; cost: number }[]; } export interface GeneratedBusinessPlan { uiLanguage?: "en" | "bn"; businessPlan: BusinessPlan; actionSteps: ActionStep[]; riskAlerts: RiskAlert[]; aiConfidenceScore: number; reasoningSummary: string; references: string[]; unitEconomics: UnitEconomics; personalizedStrategy: string; deepDives?: Record; crisisLogs?: { issue: string; advice: string; date: string }[]; financialLogs?: { date: string; feedback: string; advice: string }[]; warRoomLogs?: { date: string; crisis: string; cfo: { name: string; argument: string }; cmo: { name: string; argument: string }; ceo: { name: string; verdict: string }; }[]; fundingPitch?: string; marketingAssets?: { branding?: { brandName: string; bio: string; profilePicPrompt: string; coverPicPrompt: string; brandCharacter?: { imageUrl: string; description: string }; }; contentHistory?: { date: string; type: "post" | "reels"; duration?: number; imagePrompt?: string; sequences?: { timestamp: string; cinematicPrompt: string; dialogue: string }[]; baseImage?: string; caption: string; }[]; } | null; roleplayLogs?: { scenario: string; userReply: string; aiFeedback: string; score: number; nextScenario?: string }[]; radarForecast?: { month: string; trendLevel: "High" | "Medium" | "Low"; insight: string; preEmptiveTask: string; }[]; radarSearchHistory?: { query: string; date: string; marketSummary: string; competitors: { name: string; detail: string; link?: string }[]; sources: { title: string; url: string }[]; }[]; survivalPlan?: { stepTitle: string; action: string }[]; legalDocs?: { docType: string; content: string }[]; } export interface GeneratedPlanStreamResult { data: GeneratedBusinessPlan; planId: string; } ``` ### src/app/login/page.tsx ```typescript import AuthPanel from "@/components/AuthPanel"; export default function LoginPage() { return ; } ``` ### src/app/signup/page.tsx ```typescript import AuthPanel from "@/components/AuthPanel"; export default function SignupPage() { return ; } ``` ### src/app/consultation/page.tsx ```typescript "use client"; import { FormEvent, useEffect, useState } from "react"; import { useRouter } from "next/navigation"; import { ArrowRight, BrainCircuit, CheckCircle2, Sparkles } from "lucide-react"; import { Button } from "@/components/ui/button"; type BusinessSetup = { idea: string; budget: string; language: "en" | "bn"; timeline: string; answers?: string[] }; export default function ConsultationPage() { const router = useRouter(); const [setup, setSetup] = useState(null); const [questions, setQuestions] = useState([]); const [answers, setAnswers] = useState(["", "", ""]); const [error, setError] = useState(""); useEffect(() => { let parsed: BusinessSetup | null = null; try { parsed = JSON.parse(sessionStorage.getItem("biz_setup") ?? "null") as BusinessSetup | null; } catch { parsed = null; } if (!parsed?.idea || !parsed.budget) { router.replace("/"); return; } setSetup(parsed); const { idea, budget, language } = parsed; if (!idea || !budget) { router.replace("/"); return; } let active = true; async function loadQuestions() { try { const response = await fetch("/api/questions", { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ idea, budget: Number(budget), language }), }); const result = await response.json(); if (!response.ok || !Array.isArray(result.questions) || result.questions.length !== 3) { throw new Error("Unable to prepare your consultation."); } if (active) setQuestions(result.questions); } catch (consultationError) { if (active) setError(consultationError instanceof Error ? consultationError.message : "Unable to prepare your consultation."); } } loadQuestions(); return () => { active = false; }; }, [router]); function handleSubmit(event: FormEvent) { event.preventDefault(); if (answers.some((answer) => !answer.trim())) return; if (!setup) return; sessionStorage.setItem("biz_setup", JSON.stringify({ ...setup, answers: answers.map((answer) => answer.trim()) })); sessionStorage.setItem("micromentor_language", setup.language); router.push("/dashboard"); } const language = setup?.language ?? "en"; if (!questions.length && !error) { return (

Adaptive consultation

{language === "bn" ? "মেন্টর আপনার ধারণা বিশ্লেষণ করছেন..." : "Mentor is analyzing your idea..."}

{language === "bn" ? "আপনার জন্য তিনটি নির্দিষ্ট প্রশ্ন প্রস্তুত করা হচ্ছে।" : "Preparing three precise questions for your business reality."}

); } if (error) { return

{error}

; } return (

Super-human mentor consultation

{language === "bn" ? "আপনার বাস্তবতা বুঝতে চাই" : "Let’s understand your real operating reality"}

{language === "bn" ? "আপনার উত্তরগুলো দিয়ে এআই পরিকল্পনা, ঝুঁকি ও ইউনিট ইকোনমিকস ব্যক্তিগতভাবে সাজাবে।" : "Your answers shape the strategy, risks, and unit economics around you."}

{questions.map((question, index) =>