Video summary

How I Built a $50,000 Competitor Spy Tool with Claude Code (Tutorial)

Main summary

Key takeaways

Technology

Summary of the Video’s Technological Content (Tutorial)

Problem Being Solved

The tutorial focuses on turning “AI everywhere” into practical business value, specifically for:

  • Market intelligence
  • Competitor ad analysis

The goal is to save hours of manual research by automating competitor monitoring and analysis.

Core Idea (Product Concept)

The speaker proposes building a competitor spy / market intelligence tool that:

  • Scrapes competitor ad data from multiple channels
  • Uses an LLM workflow to analyze the scraped data
  • Produces a dashboard for fast exploration of competitor activity, including:
    • Ad formats
    • Volume
    • Copy
    • CTAs
    • Impressions, and related signals

The tutorial claims the approach can replace paying $200–$500/month for similar tools.

Tools Used

  • Claude Code
    • Used for LLM-assisted coding and “research mode”
    • Generates a detailed specification, then writes the full multi-page dashboard application
  • API as a scraping layer
    • Used to scrape ad/library data from sources such as:
      • TikTok
      • Google Maps
      • Instagram
      • LinkedIn (used in the tutorial example)

In the example, the scraping target is specifically LinkedIn ads.

Workflow Overview (Step-by-Step)

  1. Choose a business context/persona

    • Example: a London marketing agency doing SEO + paid social
    • Mentions that prompts may also be oriented toward e-commerce/DTC depending on the use case

  2. Prompt Claude Code for market research

    • Claude Code is instructed to:
      • Use research mode
      • Generate a competitor list
    • Optionally improves results by providing context via uploads (e.g., website/decks)
    • Claude Code can also ask follow-up questions automatically (via “recent updates”)

  3. Extract competitor identities

    • Claude outputs competitor names/identities (examples mentioned: “key marketing”, “the good marketer”)

  4. Scrape LinkedIn ad data

    • Uses a third-party LinkedIn ads scraper (example mentioned: by Gustavo Silva Cohorent)
    • Steps include:
      • Add competitor accounts into the scraper’s bulk edit
      • Set a large scrape count to capture “everything”
      • Run the scrape
    • The tutorial notes you can test with a small free credit (e.g., $5)

  5. Export the scraped dataset

    • Exported from the scraper as JSON

  6. Feed the dataset back into Claude Code

    • Attach/export the JSON dataset to provide maximum context
    • Key principle emphasized: longer prompts / more context → better results
    • Claude Code then generates a large output described as a ~5,000-word specification (and/or a 918-line spec)

  7. Generate and host the dashboard application

    • Claude Code creates an application codebase described as multiple pages (e.g., “Dashboard”, “Library”, etc.)
    • Hosting options mentioned:
      • Netlify (used in the tutorial)
      • Vercel (alternate)

Netlify deployment flow:

  - Import an existing project from GitHub into Netlify
  - Claude Code pushes changes to a **secondary branch**
  - Create a **pull request** and merge into **main**
  - Netlify deploys the site
  1. Fix runtime issues if needed
    • If the site loads blank/white:
      • Check the browser console (developer tools)
    • Then instruct Claude Code to fix the issue and re-run build/deploy

Final Product Features Showcased (Dashboard Capabilities)

The dashboard includes:

  • Total ads tracked
  • Active competitors
  • Ad formats detected
  • Ads with impression data
  • A searchable ad library with:
    • Filtering by company
    • Filtering by format (e.g., video ads)
  • A way to view the competitor LinkedIn page directly from the dashboard

Analytics / Competitive Landscape Views

The dashboard supports analysis such as:

  • Ad volume comparison
  • Format strategy radar
  • Body copy / headline length
  • Content intelligence
  • CTA analysis and a “diverse score”
  • Impression intelligence

The speaker claims the structure is reusable: the tool can be adapted to different companies and contexts by adjusting:

  • prompts
  • scraping targets

Review / Guide Nature

This is primarily a hands-on implementation tutorial, including:

  • Step-by-step setup
  • Configuration advice (scraper selection + credentials)
  • Deployment guidance (Netlify)
  • Debugging steps (use dev tools/console and ask Claude Code to fix issues)

Main Speakers / Sources

  • Main speaker: Riad
    • Works in revops and digital growth
    • Implements AI strategies and presents the tutorial
  • Tools / sources referenced:
    • Claude Code (Anthropic)
    • External LinkedIn ads scraper by Gustavo Silva Cohorent
  • Hosting platforms referenced:
    • Netlify (primary)
    • Vercel (alternate)

Original video