# How to Vibe Code Your Own LLM Pulse (and Stop Paying for It)

> All-in-one AI search visibility & reputation platform

- Site: https://llmpulse.ai
- Category: Marketing & SEO SaaS
- Verdict: **Serious undertaking** (45/100 vibecodeable)
- Estimated effort: 4-6 weeks of part-time work

## Verdict

Build a personal prompt tracker clone, but expect significant engineering pain maintaining reliable scrapers against shifting AI UIs.

The dashboard and database CRUD are straightforward with AI coding tools, but the data ingestion pipeline will break constantly. Querying multiple live AI engines at scale without triggering blocks, rate limits, or DOM changes requires resilient queuing and retry logic that consumes weeks of debugging.

### What you can't replicate

- Historical data archives of AI search volatility across millions of prompts
- Enterprise white-label agency portal infrastructure and customer trust

## What it does

Monitor, analyze, and optimize brand visibility across ChatGPT, Perplexity, Gemini, and Google AI search engines.

### Core features

- Automated prompt query scheduling across multiple AI engines
- Citation source intelligence and URL extraction
- Sentiment classification and brand reputation scoring
- Competitor share-of-voice benchmarking
- AI traffic analytics integration (GA4 / Plausible)
- Agent analytics (tracking AI crawlers like GPTBot)

## The business

### Pricing

- Starter: €49/mo — For individuals and small teams.
- Scale: €299/mo — Most popular tier for organizations.

### Funding

$0 raised.

Founded 2024.
Team size: 3.

## The hard parts

- Bypassing bot-detection and rate limits on live consumer-facing AI interfaces
- Reliably parsing messy markdown and anchor tags into canonical citation domains
- Asynchronous scheduling and orchestration of thousands of concurrent AI runs
- Maintained integrations with external web analytics and BI connectors

## How to vibe code LLM Pulse

### Prerequisites

- Node.js (Free): Runtime for Next.js full-stack application development
- GitHub (Free): Code repository and deployment source control

### Recommended AI tools

- Claude Code: Best-in-class agentic coding tool for scaffolding full-stack features and background workers from the terminal
- Cursor: AI-native code editor for fine-tuning dashboard views and UI components

### Stack

- Frontend: Next.js
- Backend: Next.js Server Actions / API Routes
- Database: Neon
- Auth: better-auth
- Payments: None
- Other: Vercel AI SDK, Inngest, PostHog

### Hosting

- Vercel (Hosting Next.js frontend and serverless API functions): $0-20/mo
- Neon (Serverless Postgres database for projects, prompts, and response histories): $0/mo

### Build guide

1. **Project Scaffolding & Database Schema** — Initialize the Next.js application, configure Tailwind CSS, set up better-auth, and define the Neon database schema for projects, prompts, responses, and competitors.

```
Scaffold a new Next.js project with Tailwind CSS and TypeScript. Integrate better-auth with email/password authentication connected to a Neon PostgreSQL database using Drizzle ORM. Define database tables for `projects`, `prompts`, `competitors`, `runs`, and `responses`. Each run should link a prompt to an AI model, storing the raw response text, extracted brand mentions, sentiment score (-1 to 1), and JSON-serialized citation URLs. Ensure clean separation of schema files and proper foreign key constraints.
```

2. **Dashboard UI & Project Management** — Build the core dashboard interface for managing projects, adding tracked prompts, and setting up competitor lists.

```
Create a responsive dashboard layout in Next.js with a sidebar navigation and top project selector. Implement pages for managing projects, a prompt tracking table with pagination and filtering, and a competitor management modal. Use shadcn/ui components for tables, dialogs, forms, and status badges. Connect the UI to server actions that perform CRUD operations on projects, prompts, and competitors in the Neon database.
```

3. **AI Model Orchestration & Response Parsing** — Implement the core ingestion pipeline that submits prompts to target AI models via API, extracts brand mentions, parses citation URLs, and scores sentiment.

```
Build a backend service module using the Vercel AI SDK and direct API clients for OpenAI (ChatGPT), Anthropic (Claude), and Google Gemini. The service must accept a prompt and target model, execute the query, and return the raw text along with search grounding metadata or markdown anchor links. Implement a parsing utility that extracts all citation URLs, normalizes them to canonical domains, calculates share of voice, and runs a lightweight LLM call to classify brand sentiment on a 5-point scale. Store the results in the `runs` and `responses` tables.
```

4. **Asynchronous Scheduling & Background Jobs** — Set up Inngest to handle weekly and scheduled recurring prompt runs across multiple AI engines reliably without timing out serverless functions.

```
Integrate Inngest into the Next.js app to handle background job execution. Create a durable background workflow function that triggers weekly (or on demand), iterates over all active projects and their tracked prompts, fans out execution across configured AI models using the orchestration service from step 3, and saves results to the database. Implement robust error handling, rate-limit backoff logic, and step-level retries for each AI provider call.
```

5. **Analytics & Visualization Views** — Build charts and reporting views to display brand visibility scores, share of voice over time, citation sources, and sentiment breakdowns.

```
Create a comprehensive analytics view for projects using Recharts. Include line charts showing brand visibility and sentiment trends over time, bar charts for competitor share of voice comparison, and a data table summarizing top trusted citation domains and their occurrence counts. Add date-range filters and export-to-CSV functionality for the reporting data.
```

### Cost vs paying

**Starting costs (one-time):**

- Custom domain: $12/yr
- Total: ~$12 one-time

**Ongoing costs (monthly):**

- AI API Usage (OpenAI/Anthropic/Gemini for prompt tracking): ~$15-30/mo
- Vercel & Neon Hosting: $0-20/mo
- Total: ~$30-50/mo

- Paying for the SaaS instead: €299/mo (Scale tier)
- Build time: 40-60 hours
- AI tool credits: $20/mo (Claude Pro)
- Break-even: 1 month

## Sources

- [LLM Pulse Official Website](https://llmpulse.ai)
- [LLM Pulse Pricing](https://llmpulse.ai/pricing)