Competitive intel for people who don't have a growth team
REM turns one URL into a cited teardown of a company's entire marketing machine — ads, SEO, founders, funnel and all.
What it does
You give it a company’s URL and it scrapes every public trace of their growth strategy: Meta/Google/LinkedIn ad libraries (including paused campaigns and the longest-running winners), tracking pixels hidden inside GTM containers, sitemap structure to spot programmatic SEO families, Wayback Machine positioning history, founder identities, Reddit activity, and app store data. The output is a written playbook with every claim tied to an evidence citation — and an explicit list of what it couldn’t see, so you know where the blind spots are.
The interesting bit
The epistemics are the unusual part: unknowns are surfaced rather than papered over, and the MCP server includes a validate_report tool that checks the agent’s report for invented spend figures and unresolvable citations. It also runs the report-writing through your existing Claude Code or Codex subscription (with API keys stripped from the environment) instead of billing tokens.
Key highlights
- Works key-free for a surprising amount: pixels, sitemaps, Wayback history, App Store data, HN launches, funnel analysis
rem diffre-runs a company and reports deltas — new/killed ads, follower growth, repositioning- MCP server with guided prompts like
compare_competitorsandsteal_the_playbook(teardown → 30-day plan for your product) - Local web app with a review library of every ad, post and video collected, sortable by impressions or run length
- Apple Search Ads forensics from the EU repository, including copycat apps bidding on the brand name
Caveats
- The juicy layers (ad libraries, social stats, Reddit, founder enrichment) need a paid Scrape Creators key, at roughly 25–50 credits per run
- The written strategy report requires either an Anthropic API key or a Claude/ChatGPT subscription with the CLI installed
- The README is openly a funnel for the author’s UGC ad-clip service — fitting, given the subject matter
Verdict
Genuinely useful for founders and marketers doing competitive research on a budget, and a nice example of evidence-cited agent tooling. If you don’t care about marketing, there’s nothing here for you.
Frequently asked
- What is dansugc/reverse-engineer-marketing?
- REM turns one URL into a cited teardown of a company's entire marketing machine — ads, SEO, founders, funnel and all.
- Is reverse-engineer-marketing open source?
- Yes — dansugc/reverse-engineer-marketing is open source, released under the MIT license.
- What language is reverse-engineer-marketing written in?
- dansugc/reverse-engineer-marketing is primarily written in TypeScript.
- How popular is reverse-engineer-marketing?
- dansugc/reverse-engineer-marketing has 1k stars on GitHub.
- Where can I find reverse-engineer-marketing?
- dansugc/reverse-engineer-marketing is on GitHub at https://github.com/dansugc/reverse-engineer-marketing.