This changelog highlights releases that materially changed LinkedIn Company Scraper for users. Dates use UTC; small internal refactors and routine maintenance are intentionally omitted.
Fixes
- Company names of big brands such as Stripe and Microsoft now resolve. LinkedIn lists several pages under the same name, and these were treated as an ambiguous match and skipped.
Fixes
- Old URLs of renamed companies that LinkedIn forwards to a regional site (for example th.linkedin.com) now return the company instead of failing after three attempts.
- A run no longer hangs at start when the platform is slow to confirm its error-reporting setup.
Fixes
- A request through a proxy that never answers no longer hangs until the run times out: every request now ends within its time limit, so a short run timeout finishes with its results and a clear status.
- A company page that LinkedIn blocks on ordinary requests is now named as such in the status message, instead of "try them again later": retrying does not open it until the premium route is available. It still costs nothing.
Pricing
- From November 20, 2026 the start fee drops from $0.005 to $0.001, and the per-company price becomes FREE $0.003, BRONZE $0.003, SILVER $0.0027, GOLD, PLATINUM and DIAMOND $0.0024.
- Planned from December 22, 2026: a new charge for a company page that LinkedIn blocks on ordinary requests and that a premium proxy opened (FREE $0.001, BRONZE $0.001, SILVER $0.0009, GOLD, PLATINUM and DIAMOND $0.0008), only when that company is saved. Ordinary companies never pay it.
Features
- New fields from the same page, at no extra cost:
companyId, tagline, canonicalUrl, universalName, companySizeMin, companySizeMax, locations (every office), similarCompanies and affiliatedPages.
maxCompanies now defaults to 1,000 and goes up to 10,000, and the default run timeout is 1 hour, so long lists finish without extra settings.
Fixes
- Many renamed companies now return data: LinkedIn answered such URLs with a block (999) on every plain request, so they came back empty. From December 22, 2026 a premium route is planned to open the ones that stay blocked, and pages moved to
/school/.
- Regional URLs (
de.linkedin.com/company/...) now return all fields in English, and sub-page URLs (/about/, /jobs/) read the company's main page.
- A run that returns no companies is no longer charged the start fee. Empty input, wrong slugs and pages that could not be loaded cost nothing.
- A spending limit (
maxTotalChargeUsd) now stops the run cleanly: no rows beyond the limit, and no pages fetched that the limit could not pay for, the premium route included.
- If a result cannot be saved or a charge cannot be confirmed, the run now stops and is marked as failed, with a clear status, instead of carrying on.
- Security: certificate checks now apply to every request except the one proxy route that cannot use them, and redirects are followed only to public LinkedIn company, school or showcase pages.
- The final status message now says how many companies you got out of how many you asked for, and why any are missing. Wrong slugs stop after one quick re-check instead of three tries.
Fixes
- Large batches no longer run out of memory at the default 256 MB: each company page now uses much less memory while it is read, with the same output fields.
- LinkedIn login pages are never saved as companies. Numeric company IDs (
linkedin.com/company/1441) need login on LinkedIn; they are now skipped quickly without a charge, and the status message asks for the company page URL with its name instead. Previously such inputs could produce a "Sign in" row.
- Person profile URLs and other links that are not LinkedIn company pages are skipped with a warning instead of being fetched.
- The
website field is now always a full URL; some companies list a bare domain such as example.com, which is now returned as https://example.com.
- Company names are now looked up and scraped in parallel, so long name lists finish within the default run timeout far more often.
- When a company name list reaches
maxCompanies, the names earliest in your list always get the places, whichever lookup finishes first. Pages that do not identify themselves as a LinkedIn company page are never saved or charged.
- A name list that has already filled
maxCompanies is no longer marked as failed when a lookup for a later, unneeded name runs out of time.
Fixes
- Scraped companies are now saved and charged as they finish instead of only at the end of the run, so a run that times out keeps everything it already scraped.
- Runs now stop starting new companies shortly before the run timeout, keep everything already scraped, and end as failed with a status message and a
RUN_SUMMARY record of scraped, unfinished and not-started companies. Exact company URLs are scraped before company names are looked up.
- API inputs that send company URLs as
{ "url": ... } objects are now accepted; blank or unsupported list items are skipped with a warning instead of failing the whole run.
- Added private error reporting for failed runs and operations; see the README FAQ for what is sent.
Features
- Added bounded parallel company-page scraping with a configurable
maxConcurrency setting (default 3, range 1–10) for faster bulk runs while preserving input order and existing retry behavior.
Features
- Added optional
companyNames resolution for users who do not know exact LinkedIn slugs. Resolved rows include the submitted name, confidence, and source; ambiguous matches are skipped.
Features
- Added public Store task examples so users can start common LinkedIn company scraping workflows faster.
Integrations
- Improved schema metadata for automation and MCP clients with cleaner field titles and descriptions.
Features
- Initial release for scraping LinkedIn company profile data, including website, industry, company size, headquarters, specialties, employee count, and profile metadata.