Review AI Analyzer

(WIP)

Description

Welcome to the Review AI Analyzer! Share your product or topic reviews, and our AI will provide summaries of the good, bad, and areas for improvement.

Live Demo 🚀

Key Features:

• Authentication.
• Create Entries/Topics for review.
• Create and delete Reviews about that entry/topic.
• When creating a new review, AI will automatically make a summary about good, bad and improvement points

Technology

• NextJS;
• Typescript;
• Tailwind;
• Prisma;
• MySQL;
• PlanetScale;

To Do:

• Add score input into review prompt;
• Add most used positive, bad, and improvements from each entry using memory model;
• Add chart to follow review progress along time;
• Complete test implementation

Setup locally

Clone repository:

git clone https://github.com/dianacpg/review-analyzer-ai.git

Install dependencies:

npm i

Create Clerk account and add all secrets in .env.local:

NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=XXXXXXXX
CLERK_SECRET_KEY=XXXXXX
NEXT_PUBLIC_CLERK_SIGN_IN_URL=/sign-in
NEXT_PUBLIC_CLERK_SIGN_UP_URL=/sign-up
NEXT_PUBLIC_CLERK_AFTER_SIGN_IN_URL=/entries
NEXT_PUBLIC_CLERK_AFTER_SIGN_UP_URL=/new-user

Create OpenAI account and add secret in .env.local:

OPENAI_API_KEY=XXXXXXX

Database was created using PlanetScale, MySQL-compatible serverless database, that has a great developer experience. To do that, create an account, a database, install planetscale CLI and connect by running:

pscale auth login

pscale connect <YOUR_DATABASE_NAME> <YOUR_BRANCH> — port 3309

,

Connect db with prisma by adding secret in .env as in planetscale documentation

DATABASE_URL = 'mysql://[email protected]:3309/<YOUR_DATABASE_NAME>'

Run server

npm run dev