GRIP — Guest Review Intelligence Platform

AI-powered review management for hotels. GRIP ingests guest reviews, runs multi-dimensional sentiment analysis via Claude, auto-generates on-brand management responses, and surfaces cross-property operational insights — replacing manual review workflows with a structured, auditable pipeline.

Live demo: grip-blush.vercel.app

Try the demo

Log in at grip-blush.vercel.app with the seeded admin account:

Email	Password
demo@grip.ai	demo1234

The demo organization comes pre-seeded with multiple properties, analyzed reviews, brand voice profiles, and root cause insights.

---

Features

Area	What it does
AI Analysis	Sentiment scoring (-1→1), aspect extraction (cleanliness, service, F&B, etc.), entity recognition, urgency triage (IMMEDIATE / MONITOR / INFORMATIONAL), multilingual detection
AI Response Generation	Personalized management responses using Claude Sonnet, shaped by configurable brand voice profiles
Brand Voice Profiles	Per-property tone, personality, greeting/closing styles, do/don't phrase lists, example responses, and contextual property highlights
Bulk Approve	Review and approve multiple AI-drafted responses in one pass from the property page
Tone Override	Per-response tone dropdown to override the default brand voice at review time
Root Cause Analysis	Cross-review pattern detection that groups recurring issues, estimates frequency and revenue impact, and tracks resolution status
Portfolio Dashboard	Sentiment trends, response rate metrics, aspect radar charts, and review timelines across all properties
Multi-tenant RBAC	Organization-scoped data with four roles: Admin, Portfolio Manager, Property Manager, Responder
Audit Log	Every action (generate, approve, publish, edit) is recorded per user

---

Tech Stack

Layer	Technology
Framework	Next.js 14 (App Router, Server Components)
Language	TypeScript 5 (strict)
Database	Turso (libsql) in production · SQLite locally
ORM	Prisma 7
AI	Claude Sonnet 4 via @anthropic-ai/sdk
Auth	NextAuth.js v4 (credentials + session)
UI	Radix UI primitives + Tailwind CSS
Charts	Recharts
Testing	Jest + Testing Library — 317 tests across 13 suites
Deployment	Vercel

---

Architecture

Review ingestion
    │
    ▼
AI Analysis (Claude)         ← sentiment · aspects · entities · urgency
    │
    ▼
AI Response Draft (Claude)   ← brand voice profile resolved per property → org → fallback
    │
    ▼
Staff review + approve       ← bulk approve, tone override, edit distance tracked
    │
    ▼
Publish                      ← audit logged, response metrics updated
    │
    ▼
Root Cause Analysis (Claude) ← cross-review patterns, severity, revenue impact estimate

Project layout

src/
├── app/
│   ├── (auth)/login/            # Login page
│   ├── (dashboard)/             # Main application
│   │   ├── page.tsx             # Portfolio dashboard: metrics + timeline
│   │   ├── reviews/             # Review list + detail pages
│   │   ├── properties/          # Property pages with bulk approve
│   │   ├── intelligence/        # Sentiment trends, aspects, root causes
│   │   └── settings/            # Brand voice, users, properties
│   └── api/                     # REST API routes
│       ├── analytics/           # Portfolio summary, sentiment trends, aspects
│       ├── reviews/             # Review CRUD · analyze · generate · approve
│       ├── properties/          # Property CRUD · batch-generate · bulk-approve
│       ├── insights/            # Root cause analysis
│       └── voice-profiles/      # Brand voice CRUD
├── components/
│   ├── ui/                      # Radix-based primitives
│   ├── dashboard/               # MetricCard, PropertyCard, TimelineChart
│   ├── reviews/                 # ReviewCard, ResponseEditor, SentimentBadge
│   ├── intelligence/            # AspectRadar, SentimentChart, TrendLine
│   └── settings/                # BrandVoiceEditor, PropertyForm
├── lib/
│   ├── ai/                      # Claude integration (analyze, generate, root-cause)
│   ├── prisma.ts                # Database client (Turso prod / SQLite local)
│   ├── auth.ts                  # NextAuth config
│   ├── constants.ts             # Aspect categories, entity types, model IDs
│   └── voice-profile-resolver.ts
└── __tests__/                   # Jest tests (API, AI modules, components)

prisma/
├── schema.prisma                # Full data model
└── seed.ts                      # Demo data seeder

---

Running locally

Prerequisites

• Node.js 18+
• An Anthropic API key

Setup

# 1. Install dependencies
npm install

# 2. Configure environment
cp .env.example .env
# Set ANTHROPIC_API_KEY and NEXTAUTH_SECRET in .env

# 3. Generate Prisma client
npm run db:generate

# 4. Seed demo data (properties, reviews, users, brand voices)
npm run db:seed

# 5. Start the dev server
npm run dev

Open http://localhost:3000 and log in with the seeded credentials.

Environment variables

Variable	Required	Description
DATABASE_URL	Yes (local)	SQLite path — file:./prisma/dev.db
TURSO_DATABASE_URL	Yes (prod)	Turso database URL
TURSO_AUTH_TOKEN	Yes (prod)	Turso auth token
NEXTAUTH_SECRET	Yes	Random secret for session encryption
NEXTAUTH_URL	Yes	App URL — http://localhost:3000 locally
ANTHROPIC_API_KEY	Yes	Claude API key for all AI features

Scripts

npm run dev               # Development server
npm run build             # Production build
npm test                  # Run all tests
npm run test:coverage     # Tests with coverage report
npm run db:seed           # Seed demo data
npm run db:studio         # Open Prisma Studio
npm run db:migrate        # Run database migrations

---

Deployment

Deployed on Vercel with Turso as the production database. Set the production environment variables in your Vercel project settings, then:

vercel --prod