DEV Community

alapha888
alapha888

Posted on

I checked 85 browser-extension ideas against real store numbers before building one — the gate that killed 81 of them

Since November 15 I have been building small local-first browser extensions. Before I write a line of code for an idea, it has to pass a demand gate: at least two kinds of hard evidence. One, a competitor's real traction numbers — store user counts, rating counts. Two, search or community demand signals. Three, paid evidence — real pricing and signs of actual sales. "It can be built" and "the store is free to publish on" are supply-side facts. They do not count as demand.

So far I have checked 85 ideas in 17 batches of five. 81 are closed. Four passed the gate and got built.

The four that passed:

  • FocusBlock, a local-first website blocker. BlockSite has about 1 million users, 4.5 stars from 32.5K ratings, and charges $49.90 lifetime or $33.50 a year for Premium. Its privacy policy (July 2026) says it collects browsing data and discloses it to affiliates. The gap was a blocker that stores everything locally and makes no network requests. It is live on Edge Add-ons: https://microsoftedge.microsoft.com/addons/detail/focusblock/kdpeddhiegpfadpjiaieoaeiloebgdhp — privacy page: https://alapha888.github.io/focus-block-privacy/
  • ConvoExport, export of a single ChatGPT or Claude conversation. AI Toolbox has 40,000+ users and charges $9.99 a month or $99 lifetime per module.
  • YouTube transcript export. Transcript for YouTube has 40,000 users; Glasp has about 2 million users and a $15 a month Pro tier.
  • SessionSnap, a cross-browser session manager. Session Buddy has 1 million users and 3.3 stars from 25.1K ratings, its rating fell from 4.7, and it has not been updated since April 28, 2026. There was no session manager with an established user base covering more than one browser store.

Three of the kills, with the numbers that killed them:

  1. A local read-later replacement after Pocket shut down on July 8, 2025. The shutdown was real, but the gap had already closed: Pocket import is a free built-in in Raindrop, Readwise, Wallabag, Instapaper and others. The paid selling point I had in mind — full-text search plus Pocket import — collides head-on with Raindrop Pro at about $28 a year, which already includes import for free.

  2. A bilingual subtitle helper for Bilibili. I tested 13 real videos anonymously: zero subtitle tracks, including a positive control whose embedded page state confirmed two tracks existed. Five independent implementations agree the subtitle URL only returns data under a logged-in session. And only 1 of 12 English-learning videos had a subtitle track at all. A product that needs a login session for a feature most videos do not have is not worth building.

  3. Lightweight multi-account switching. Chrome's cookies API reference (updated October 29, 2026) confirms Chrome has only a default store and a privacy store — no API to create a new store, no per-tab store. The only pure-extension route is swapping cookies in and out, which is the same route that broke SessionBox's MV3 version, and it cannot move localStorage or IndexedDB login state. Firefox has real isolation through contextual identities, but Mozilla's own free Multi-Account Containers already covers the core need.

One more, because it shows the gate working in the other direction: hover-to-zoom images. The original HoverZoom was removed from the store for malware. The community already replaced it with a free open-source fork, Hover Zoom+, plus several free clones. No paid evidence, no gap.

What the gate does not do: it does not predict sales. FocusBlock is live; the other three were still in review at my last check this morning. Revenue so far is zero. The gate filters ideas before I spend days building them. It does not create demand — that part is still unsolved, and it is the part I am working on now, one published extension at a time.

Top comments (0)