Ohio SOS Business Registry Scraper

Role
Sole developer
Built with
Type
Data-scraping Chrome extension
Ohio SOS Business Registry Scraper, cover

Screens

  • Ohio SOS Business Registry Scraper, screen 1 of 3
  • Ohio SOS Business Registry Scraper, screen 3 of 3
  • Ohio SOS Business Registry Scraper, screen 2 of 3

A Manifest V3 Chrome extension that collects business-entity records from the Ohio Secretary of State search portal, with two scraping engines, resume-anywhere state and rate-limit recovery for unattended multi-hour runs.

Role: Sole developer, standalone build, vanilla JavaScript with no framework.

Built a Manifest V3 Chrome extension that harvests business-entity records from the Ohio Secretary of State Business Search portal. It runs as a side-panel app and offers two independent scraping engines, a DOM-automation sweep and a far faster direct-API sweep, with alphabetical prefix coverage, date and relevance filtering, and full rate-limit recovery so an unattended multi-hour run survives throttling and browser hiccups.

What I delivered

  • Dual scraping engines: a DOM engine that drives the site's DataTables search UI, and an API engine that hooks the page's XMLHttpRequest and fetch to capture the real search endpoint, then replays it per prefix on the same origin, about ten times faster.
  • Exhaustive alphabetical sweep: 1,296 two-character prefixes (AA to 99) to page past the site's result cap and reach every entity, with a live progress bar and per-prefix streaming.
  • Rate-limit (429) recovery: detects both readable and CORS-stripped 429s, reloads the page to refresh clearance, and resumes from the exact prefix it stopped on, with escalating back-off only when reloads stop making progress.
  • Proactive throttle avoidance: optional page refresh every N requests, plus randomised inter-request jitter of 1.8 to 3.5 seconds.
  • Resume anywhere: the current prefix is continuously persisted to chrome.storage.local, so a crashed, stopped or blocked sweep restarts where it left off without re-collecting.
  • CAPTCHA and bot-wall detection: halts immediately on a challenge instead of hammering the site.
  • Detail enrichment: for each new entity, fetches the detail report to pull the mailing address, with global dedup so each business is fetched exactly once.
  • Robust wait logic: a page-context waiter arms a DataTables draw listener before clicking Search and watches the loading indicator through a two-phase appear-then-vanish lifecycle, so results are never read stale or missed.
  • Transient-error retry: four rounds of exponential back-off for 5xx and malformed responses, while CAPTCHA, rate-limit and stop bubble straight up to their own recovery paths.
  • Filtering pipeline: "last N days" date window, Ohio-only relevance filter, and entity-number dedup with a name-plus-date fallback key.
  • Export: CSV and JSON of the client's exact four columns (registration number, business name, mailing address, registration date), with spreadsheet formula injection neutralised.
  • Side-panel UX: manual search, single-prefix test flow, status filters (All, Active, Cancelled, Dead, Fraudulent), a timestamped activity log, and a stop-after-current-prefix control.

Scale: 5 JavaScript modules, side-panel UI, 1,296-prefix sweep engine, about 1,800 lines across content, popup, injected, API-hook and background scripts.

Stack

  • Chrome Extension Manifest V3
  • Vanilla JavaScript
  • Chrome APIs (sidePanel, tabs, scripting, storage, downloads)
  • Page-context XHR/fetch hooking
  • DataTables / jQuery interop
  • HTML/CSS side-panel UI