AI NewsDev toolsAnnouncement

Reverse Engineer Marketing pulls a rival's whole marketing stack in one run

An MIT tool reads a rival site's ads, social, SEO and founders from public sources. Every fact cites its source.

AI News

Editorial2 min read

LinkedInX
GitHub social card for the reverse-engineer-marketing repository

Image: GitHub

Why it mattersA marketer or founder can run one command. The report is cited. The work used to take days by hand.

Reverse Engineer Marketing, called REM in its README, is an open-source command. It reads a rival site's whole marketing stack from public sources. It writes a cited report. The author, DansUGC, published it under MIT. The repository has 469 stars. It was created on 18 October 2026, so the traction is real inside a 30-day window.

What one run produces

A user types npx reverse-engineer-marketing <domain>. The tool writes a Markdown report and an evidence.json file to a local folder. Six kinds of facts land in that report. Paid ads, pulled from Meta's Ad Library, Google's Ads Transparency Center, LinkedIn Ads, and Apple's EU Ad Repository. Organic social, pulled from TikTok, Instagram, YouTube, X, LinkedIn and Reddit. Technical signals, read from the site itself: the pixels it fires, the GTM containers around them, robots.txt, llms.txt, the sitemap, and the programmatic page families inside it. Positioning history, pulled from the Wayback Machine (homepage title, h1, call-to-action, over time). Founders and the brand's history, pulled from LinkedIn, X and Wikidata. App data, pulled from the App Store and Google Play.

Evidence over claims

The README puts the point in one line. The project says every fact in the report cites the evidence it came from. Anything the tool could not see is listed as unknown, with a link to check it by hand. Each item in the report carries an evidence ID. The ID points at a row in the JSON. An unknown is marked as unknown. A link points at the source that would have had the fact. That shape matters for a review shared with a team. The reader can audit a surprising claim rather than take it on trust.

How it runs

Two modes. The CLI runs in a terminal. It uses your Claude Code or Codex plan for the parts that need a model. A --no-ai switch runs a pure deterministic version. The MCP server lets a coding agent run the same report as a tool call during an agent session. There is also a local web app (npx reverse-engineer-marketing web) for a teammate who does not want a terminal.

Why the audience is wider than developers

Competitor research is the same work on every team. A founder, a marketer, a product manager, an agency, an investor doing diligence: they all do it. The job has always been slow and manual. The output has always been a deck of screenshots. The first page usually said "not an exhaustive list". REM matches the first part of that work to a one-line command. The output is a document a team can argue with. Each sentence points at its source. MIT means it goes inside the company without a licence talk. The repository earned 469 stars in a month. That is the signal the problem was real for far more people than build software.

Source

dansugc/reverse-engineer-marketing on GitHub.

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

Share
LinkedInX
Start a project