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

> AI Ad Generator & Campaign Studio

- Site: https://marketmaster.dev
- Category: Marketing SaaS
- Verdict: **Solid side project** (62/100 vibecodeable)
- Estimated effort: 2-3 weeks of part-time development

## Verdict

Build a single-channel version with Claude Code in 2-3 weeks, but expect friction integrating 8 different social network APIs and the dual-model validation loop.

Replicating the core workflow—scraping a URL, generating a brand kit, prompting asset generation, and running a dual-model validator—is entirely feasible with a modern TypeScript stack and AI coding tools. However, integrating the live publishing graph across eight different social platforms (LinkedIn, X, Instagram, TikTok, YouTube, Facebook, Pinterest, Threads) introduces massive OAuth maintenance debt and rate-limit edge cases that will drain your weekends.

### What you can't replicate

- The exact curated original ad structure pattern library
- Production-tested social media API token renewal handling across all platforms

## What it does

An AI-powered ad generator and campaign studio that researches your website, creates multi-format ad assets, and verifies them through a dual-model compliance engine before publishing.

### Core features

- Automated multi-page web scraping and context distillation
- Brand Kit extraction (colors, logos, voice, supported claims)
- Multi-format creative generation (static ads, carousels, videos, emails)
- Dual-model compliance engine (Anthropic vs OpenAI/Google verification loop)
- Direct publishing and scheduling across 8+ social platforms via OAuth
- Recurring market-signal automations and search console feedback loop

## The business

### Pricing

- Starter: $29/mo — For founders building a focused campaign each week.
- Pro: $79/mo — For teams creating every week and monitoring the market.
- Studio: $199/mo — For teams and agencies producing visual campaigns at volume.

Founded 2026.
Team size: 1-5.

## The hard parts

- Orchestrating asynchronous dual-model validation to catch AI slop and unsupported claims
- Managing fragmented OAuth flows and rate limits across 8+ distinct social APIs
- Programmatically assembling multi-format media assets (videos, carousels, voiceovers) reliably
- Robust HTML parsing and cleaning for arbitrary landing page research extraction

## How to vibe code MarketMaster

### Prerequisites

- Node.js (free): Runtime environment for Next.js and backend scripts
- GitHub (free): Source control and deployment pipeline sync

### Recommended AI tools

- Claude Code: Executes multi-file scaffolding, manages API integrations, and handles iterative debugging across the full stack
- Cursor: Ideal for reviewing generated code diffs and hand-tuning frontend dashboard components

### Stack

- Frontend: Next.js with Tailwind CSS and shadcn/ui
- Backend: Next.js Server Actions and API Routes
- Database: Turso (SQLite at the edge)
- Auth: better-auth
- Payments: Stripe
- Other: Anthropic API, OpenAI API, Fal.ai, ElevenLabs, Firecrawl

### Hosting

- Vercel (Next.js frontend and serverless API route hosting): $0-20/mo
- Cloudflare (Edge caching and object storage for generated media assets via R2): $0/mo

### Build guide

1. **Project Scaffolding and Database Schema** — Initialize a Next.js project with Tailwind CSS and shadcn/ui. Set up Turso for SQLite storage and configure better-auth for user authentication.

```
Create a new Next.js project using App Router, Tailwind CSS, and TypeScript. Install better-auth configured with Turso (libSQL) as the database adapter. Create database schema migrations for Users, BrandKits, Campaigns, CampaignAssets, and PublishLogs tables. Add basic user session protection middleware for dashboard routes.
```

2. **Website Scraping and Brand Kit Extraction** — Implement the source research step by scraping a target URL, extracting text content, logos, and color palettes, and structuring them into a reusable Brand Kit.

```
Build an API route that accepts a target website URL. Use Firecrawl or native fetch/cheerio to scrape the homepage, pricing, and product pages. Pass the extracted raw text to OpenAI GPT-4o with a structured JSON output prompt to extract audience description, offer details, differentiation points, proof items, color hex codes, and logo image URLs. Save these results into the BrandKits table linked to the user.
```

3. **Campaign Briefing and Ad Pattern Selection** — Build the campaign goal selector and verified ad structure pattern selector UI and database mapper.

```
Create a multi-step campaign creation wizard UI in Next.js. Step 1: Input target URL and display extracted Brand Kit for user review and edits. Step 2: Select campaign goal (lead gen, product launch, awareness) and choose from a predefined list of 5 verified original-ad pattern structures (e.g., Problem-Agitate-Solve, Before-After-Bridge). Save brief state in local workflow state before triggering asset generation.
```

4. **Multi-Format Asset Generation Pipeline** — Connect AI text and media models to generate static ad copy, carousel slides, short script ideas, and campaign emails based on the approved Brand Kit and ad pattern.

```
Build a campaign generation background action that takes the approved Brand Kit and Ad Pattern. Call Anthropic Claude Sonnet to generate platform-specific social posts (LinkedIn, X, Instagram), a 5-slide carousel text structure, and a campaign email. Simultaneously call Fal.ai API for static ad image generation using Flux Schnell, incorporating brand colors. Store all generated assets in the CampaignAssets table with status 'pending_review'.
```

5. **Dual-Model Evidence Engine Verification** — Implement the compliance review step where two independent AI models review generated assets for claim safety, hook strength, and AI slop.

```
Implement the Evidence Engine validation loop. When assets are generated, send them in parallel to two evaluation prompts: one evaluated by Anthropic Claude Sonnet and one by OpenAI GPT-4o. Each model must output a strict JSON payload with boolean approval, list of flagged slop phrases, and claim safety score against the source Brand Kit evidence. Require both models to return approved: true before unlocking the publication UI. Render clear audit feedback badges on each asset card in the dashboard.
```

6. **Publishing, Scheduling, and History Tracking** — Build external API connectors and UI to schedule or immediately post verified assets to connected social channels.

```
Build social platform publishing integrations for LinkedIn and X (Twitter) using official OAuth2 endpoints and API v2 posting routes. Create a Publish/Schedule dashboard view that lists approved campaign assets, displays scheduling selectors, and tracks live provider post IDs, permalinks, and sanitized failure logs in the PublishLogs table.
```

### Cost vs paying

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

- Domain name & developer accounts: $12 one-time
- Initial API credits (OpenAI, Anthropic, Fal.ai): $25 one-time
- Total: ~$37 one-time

**Ongoing costs (monthly):**

- Vercel hosting: $0-20/mo
- Turso database: $0/mo
- LLM & Image API usage (Anthropic, OpenAI, Fal): ~$15/mo
- Total: ~$15-35/mo

- Paying for the SaaS instead: $79/mo (Pro plan)
- Build time: 40-60 hours
- AI tool credits: $20 (Claude Pro)
- Break-even: If valued at your building time, paying $79/mo for Pro is far cheaper than 50 hours of debugging social OAuth flows.

## Sources

- [MarketMaster Official Website & Product Tour](https://marketmaster.dev)
- [MarketMaster Pricing Page](https://marketmaster.dev)
- [DanielLaunches – MarketMaster Launch Profile](https://daniellaunches.com)