sheet 01 / 04linkfetch / § about
§ about

The LinkedIn layer for agents — built the boring way.

LinkFetch ships typed LinkedIn data into Claude, Cursor, and any other tool that speaks MCP. We got here by picking an unglamorous premise: no fake accounts, no stolen sessions, no clever tricks. Just public data, fetched through the user’s own consent.
§ 02 · the why

When Proxycurl shut down in July 2025, we wrote the replacement.

For five years, the LinkedIn data industry ran on a single trick — rented accounts, rotating proxies, session tokens traded on Telegram. Then the compliance bill came due. Proxycurl closed. The market froze. Agent builders scrambled.

We looked at the problem and realised every workaround shared one mistake: pretending to bea LinkedIn user. LinkFetch doesn’t. Your user installs our Chrome extension; their session stays theirs; we fetch on their behalf. It’s slower per-call. It’s also the only model that survives the next five years of enforcement.

§ 03 · what we believe

Four rules. Every product decision goes through them.

01rule · fixed

User is the principal.

Every profile lookup happens through the end-user's own LinkedIn session, via the LinkFetch Chrome extension. No fake accounts, no session impersonation, no stolen cookies. If LinkedIn can see it, and the user can see it, LinkFetch can fetch it — for that user.

02rule · fixed

Public data only.

We don't scrape behind logins we don't own. We don't rent accounts. We don't pretend to be you. The boring answer is often the right one for compliance.

03rule · fixed

Provenance on every row.

`source`, `fetched_at`, `freshness_days` on every response. If you can't explain where your data came from, you can't ship agent workflows to regulated customers. We give you the audit trail by default.

04rule · fixed

Developers first.

Plain HTTPS and JSON, curl that just works, one-block MCP install. Pricing you can do in your head. Docs written by the person who shipped the code. No sales call to try anything.

§ 04 · the short history

How we got here.

  1. Jul 2025Proxycurl shuts down. The industry's compliance story breaks overnight.
  2. Apr 2026LinkFetch API, docs and MCP server go public: session-based, public data only, user is principal.
  3. Apr 27, 2026The jobs index goes live worldwide — every public posting across 218 regions, refreshed continuously.
  4. Aug 31, 2026The LinkedIn Jobs Dataset launches: the whole archive as a $199 snapshot or $49/month live access.
  5. NextPush delivery for new postings, more datasets, and the Chrome extension in the Web Store. See the roadmap.
§ 05 · where we stand

What we commit to.

team
Remote · async-first
data model
User is principal
compliance
GDPR + CCPA
provenance
On every row
§ 06 · brand kit

Logos + usage.

write us a letter

Got a hard compliance question? We like those.