How to vibe codeCueScout

Prerequisites

• Node.jsfree

  Required for running the Next.js framework and TypeScript tooling.

• GitHubfree

  Source control and seamless deployment integration.

• 

  Perplexity API / OpenAI APIUsage-based

  Required for querying AI visibility checks and generating content drafts.

AI coding tools

• Claude Code

  Ideal terminal agent for scaffolding the full-stack Next.js app, setting up database schemas, and writing complex scraping and parsing routines.

• Cursor

  Great for iterative frontend component development on the dashboard views and report tables.

Build guide

1. 01Project Scaffolding & Database Schema

   Initialize the Next.js project with Tailwind CSS, configure Drizzle ORM with Turso SQLite, and establish the data models for products, scans, citation sources, and writing plans.

   Scaffold a new Next.js 16 application using App Router, TypeScript, and Tailwind CSS. Initialize Drizzle ORM configured for Turso SQLite. Create database tables for 'products' (id, name, description, category, user_id), 'scans' (id, product_id, status, created_at, raw_payload), 'citation_sources' (id, scan_id, url, host, classification, verdict, rank), 'buyer_questions' (id, scan_id, question, intent_score, count), and 'writing_plans' (id, product_id, title, target_date, status). Include basic connection utilities and environment variable validations.

   ⧉copy promptopen in Claude CodeCopen in Codexopen in Cursor

2. 02AI Visibility & Engine Checker Module

   Build backend routines that prompt Perplexity and ChatGPT models with target buyer questions, capturing answers and structured citation blocks.

   Build a TypeScript service module using the Vercel AI SDK that executes visibility checks against Perplexity and OpenAI models. Given a product description and a list of generated buyer questions, query the LLMs, extract the response text, and parse the structured citation URLs returned in the metadata. Save the resulting check records and extracted citation URLs into the database linked to a scan ID. Handle API errors, timeouts, and rate limits gracefully with retry logic.

   ⧉copy promptopen in Claude CodeCopen in Codexopen in Cursor

3. 03Discussion Scraping & Intent Clustering

   Implement scraping routines for public discussion platforms (Reddit, Hacker News) or integrate Firecrawl/Exa to harvest repeating buyer intent questions.

   Implement a discussion harvesting service using Firecrawl and custom RSS/API fetchers to search public forums like Reddit and Hacker News for category keywords. Process the retrieved threads with an LLM call to extract repeating buyer questions, filter out irrelevant chit-chat, assign an intent score, and store them in the 'buyer_questions' table. Group duplicate questions and track their occurrence frequencies across scans.

   ⧉copy promptopen in Claude CodeCopen in Codexopen in Cursor

4. 04Site Readiness & Rank Overlap Checkers

   Develop deterministic site readiness analyzers to inspect robots.txt rules for AI crawlers, validate JSON-LD structured schema, and check search engine rank overlap.

   Create a deterministic site-auditing engine that fetches a target website's robots.txt to inspect crawler access rules for bots like GPTBot, ClaudeBot, and PerplexityBot. Parse the homepage HTML to validate JSON-LD structured schema (Organization, FAQPage) and check for server-side rendering versus client-side hydration. Additionally, implement a search rank overlap checker that queries Brave Search or Exa to find whether cited URLs rank in organic search results for their associated buyer questions.

   ⧉copy promptopen in Claude CodeCopen in Codexopen in Cursor

5. 05Dashboard UI & Domain Leaderboard

   Build the main analytics dashboard displaying GEO scores, domain leaderboards, citation source classifications, and scan debug breakdowns.

   Build a responsive Next.js dashboard using Tailwind CSS and shadcn/ui components. Create views for the main workspace, displaying a consolidated GEO score, visibility metrics, and a Domain Leaderboard that aggregates host citation shares. Include a Citation Sources tab with classification badges (own-it, earn-it, get-listed, competitor-owned, blocked) and a scan debugging panel showing exact queries run, sources fetched, and rejection counts.

   ⧉copy promptopen in Claude CodeCopen in Codexopen in Cursor

6. 06Writing Plan & Shareable Opportunity Reports

   Implement the 30-day content publishing plan generator and HMAC-signed login-free shareable public report pages.

   Build a content planning engine that converts clustered buyer questions into a dated 30-day writing roadmap with weekly targets and AI page draft generators using the Vercel AI SDK. Then, implement a cryptographically signed HMAC token authentication mechanism for shareable Opportunity Reports, allowing external clients or prospects to view a live, read-only report at `/report/[token]` without requiring user authentication or exposing database IDs.

   ⧉copy promptopen in Claude CodeCopen in Codexopen in Cursor

Cost vs paying for CueScout