Voci AI - Sell Project

Voci — Voice-First Personal Finance Tracker Speak your spending. AI handles the rest. Voci is a fully built, fully deployed voice-first expense and income tracker. A user speaks a transaction — "spent $50 on groceries" — and AI transcribes it, categorizes it, and matches it to the right budget automatically, with the entry appearing on their dashboard already tagged and organized. No typing, no spreadsheets, no manual categorization. Condition & Functionality This is not a prototype, mockup, or partially-built MVP — every feature described here is live, tested, and running in production right now at the app's deployed URL. The full user journey works end to end: sign up, record or manually enter a transaction, watch AI categorize and budget-match it in real time, track spending against budgets, manage a subscription plan through Stripe, and install the app to a home screen like a native mobile app. A buyer can open the link today and use the complete product exactly as a real customer would. Core Features Voice-to-transaction AI pipeline. In-browser audio recording with live waveform feedback during capture AI transcription of natural spoken language — no rigid command syntax required AI categorization across 10+ spending categories Confidence-scored AI budget matching — a transaction is only auto-linked to a budget when the AI is genuinely confident, protecting the integrity of the user's budget data instead of forcing false matches Manual entry available as a fallback or preferred alternative input method Installable Progressive Web App Installs directly to a home screen on Android, desktop, and iOS — no app store submission, review process, or approval delay required Launches full-screen like a native app once installed, with no browser address bar or chrome Custom service worker enables offline resilience and faster repeat loads Smart budgeting engine Users create budgets by category and target amount Every new transaction is automatically checked against existing budgets and linked when it genuinely fits Real-time spend tracking, remaining balance, and visual progress indicators per budget Full subscription billing infrastructure 5 pricing tiers (Free, Starter, Basic, Pro, Pro Annual) already configured Stripe Checkout integration for plan upgrades Webhook-driven plan sync handling renewals, scheduled cancellations, and failed payments Automated daily job that downgrades expired subscriptions to Free and warns users before their access lapses — a detail most MVPs at this stage skip entirely Transaction management Full transaction history with delete support Filterable by type (income/expense) and by date range (today / this week / this month) Income and expense totals calculated live Reports & premium features (plan-gated) PDF report generation, invoice creation, CSV export AI chat and AI-generated spending insights Specifications / Tech Stack Frontend: Next.js 14 (App Router), React, Tailwind CSS Backend: Next.js serverless API routes Database: Firebase Firestore, with Firebase Admin SDK for secure server-side access Auth: Firebase Authentication AI: Groq API running Llama models for transcription, categorization, and budget matching Payments: Stripe (Checkout, Subscriptions, Webhooks) PWA: Web App Manifest + custom-written service worker — no third-party PWA library or build tooling dependency Hosting: Vercel, serverless, with a scheduled Vercel Cron job handling subscription lifecycle checks The entire architecture runs on serverless infrastructure — there are no always-on servers to pay for, so hosting costs stay near zero until the app has real traffic to scale into. What Makes This Stand Out Most personal finance apps still require manual typing or, at best, a receipt photo. Voice-first competitors do exist (Finny, Money Vault, Wallet by BudgetBakers), but this product combines voice capture with confidence-gated AI budget matching — meaning it's specifically engineered not to silently corrupt a user's budget with a bad AI guess, a subtlety most competing tools don't handle carefully. Layer in a fully working multi-tier Stripe billing system with real lifecycle management (not just a one-time checkout), and a genuinely installable PWA, and this is a technically complete product — not a UI shell wrapped around a single feature. Why This Is a Great Choice Building this from zero means solving several genuinely difficult integration problems: chaining voice capture through AI transcription, AI categorization, and AI budget matching into one reliable pipeline; building correct Stripe subscription logic that handles renewals, cancellations, and failed payments without leaving users in a broken state; structuring a Firestore data model that scales cleanly across users; and making the whole thing installable as a PWA. All of that engineering is already done, tested, and running. A buyer acquires: The complete, working codebase Copyright to all original code The live, deployed application A technically s...