From the Lab

AI Competitor Toolkit

systematic competitive analysis

claude-codeplaywright-mcptemplates
$ cat problem.md

The Problem

Competitive research is one of those “important but always pushed” tasks because it’s painfully manual:

  • Click around a bunch of sites
  • Screenshot everything “just in case”
  • Try to compare pricing/features/messaging across different IA patterns
  • Lose your notes or never turn them into something shareable

I wanted a workflow that makes competitor research repeatable, fast, and organized (so the output is usable, not just a pile of tabs).

$ cat solution.md

The Solution

I built a template repo that turns competitor research into a structured project:

  • A research template that guides what to look for (pricing, positioning, UX, GTM, proof, etc.)
  • Automated browser capture with Playwright MCP
  • Faster synthesis and report writing with Claude Code
  • A consistent folder structure for screenshots, reports, and status tracking

Repo: ai-competitor-toolkit

$ cat stack.md

Technical Stack

This is intentionally lightweight — the “product” is the workflow:

  • Claude Code (CLI): drives the analysis and report generation
  • Playwright MCP: navigates pages and captures screenshots reliably
  • Node scripts (optional): research one competitor or batch through a list
  • Markdown + JSON: simple, portable artifacts that work in any repo
$ cat impact.md

Impact

Research on Rails

The biggest win is consistency:

  • Every competitor gets the same baseline review (so comparisons are real)
  • Screenshots are organized and easy to reference later
  • Reports follow a predictable structure, so you can actually use them

It also makes the task easy to delegate or hand off — you’re not relying on one person’s personal research habits.

THE RECEIPTS

Directory Structure

.
├── competitors.json          # your competitor list (create from template)
├── PROJECT-CONTEXT.md        # project briefing (create from template)
├── research-template.md      # the research checklist
├── screenshots/              # auto-captured during research
├── reports/                  # generated writeups per competitor
└── scripts/                  # optional automation

It’s designed to be copied into any product repo (or used as a standalone workspace).

Typical Run

  1. Define your baseline (your product + what you care about)
  2. Add competitors to competitors.json
  3. For each competitor:
    • capture key pages (home, pricing, features, docs, customers)
    • extract claims + differentiators
    • write a structured report
  4. Mark status + date so you can track progress over time

The goal is to make “competitive research” feel like running a checklist, not doing archaeology.

What You Get Per Competitor

  • Report: a single markdown writeup you can share internally
  • Screenshots: evidence for claims, pricing tables, feature lists, etc.
  • Notes/status: progress tracking so you can batch research without losing your place

If you want the template: ai-competitor-toolkit

← Back to the Lab