Proof point → By 2024 the question stopped being “can I build it?” and became “how does it reach everyone?” - so the extension shipped in fifteen languages.

🩻 Problem

Creators need their own media back - reels and tweet videos for repurposing - and the platforms don’t offer a download button. Browser extensions can, but scraping logic inside an extension breaks weekly.

🔨 Solution

A product line with the architecture split correctly:

Architecture Overview

  1. Thin extensions - Manifest V3 done properly: content scripts, service workers, scoped permissions.
  2. chromer-api - the heavy lifting centralized server-side: a headless-Chrome (browserless-pattern) Express backend with downloader and YouTube-shorts services - fragile logic kept where it can be fixed once, not shipped to every user’s browser. (Also one of the rare 2023 repos with a test file.)
  3. tweet-downloader-chrome (2024) - the productized one: popup UI, pinned extension key, externally_connectable handshake to a companion site (100xtools.com), and _locales translations in ~15 languages - Arabic to Japanese to Portuguese.

📜 Philosophy

Split products at the fragility line: stable UI in the client, breakable scraping behind an API you control. And treat localization as the cheapest global distribution there is.

🎓 Key learnings

  • Chrome extension architecture end to end: MV3 service workers, permissions, i18n, web↔extension messaging.
  • Operating headless Chrome as a service.
  • Productization instincts: the 2023 repos have empty READMEs; the 2024 one has fifteen languages - the polish moved to where users are.

📈 Output & impact

  • A working extension product line with real distribution engineering (localization, companion-site integration).

🌍 Why this matters

Original Tools & IP · The Founder Journey. Products live or die on distribution, and localization is the cheapest reach there is - a lesson that matters double in multilingual markets like Nigeria, where Yoruba-first users are not an afterthought but the strategy. The client/server fragility split is also how any integration-heavy platform should be designed.

🚀 Hire me

Building a browser product or wrestling a fragile integration? Let’s talk → · See also: Social Growth Infrastructure · The thesis