Proxycurl shut down. Here's a public LinkedIn data API to move to
Proxycurl closed on July 4, 2025 after a LinkedIn lawsuit. CrustAPI is a live LinkedIn data API that reads only public, logged-out pages, and the credits you buy never expire.
✓ CrustAPI covers profile, company, jobs, and people search
✓ method is public and logged out, no accounts, no logins
✓ paid credits never expire (Proxycurl's expired)
→ profiles come back in about a second (0.8s server time, tested)
I'm Dawson Young, founder of CrustAPI. I ran the calls on this page on July 15, 2026, and the receipt below is real output.
What happened to Proxycurl
Proxycurl, built by Nubela, was one of the biggest LinkedIn data APIs, around $10M in revenue by the founder's own number. In January 2025 LinkedIn filed a federal lawsuit against it. LinkedIn alleged the unauthorized creation of hundreds of thousands of fake accounts and the scraping of millions of member profiles.
On July 4, 2025 Proxycurl announced it was shutting down. The founder said a bootstrapped company could not afford to litigate against Microsoft-backed LinkedIn even if it might win, so it settled. The court-confirmed outcome was a permanent injunction requiring Proxycurl to delete LinkedIn data obtained through unauthorized means and to stop accessing LinkedIn unlawfully. In 2026 it is still closed, and the successor product does not scrape LinkedIn.
What LinkedIn went after was fake logged-in accounts, not public data collection. That is worth understanding before you pick your next tool. CrustAPI uses no accounts, no logins, and no fake profiles. It reads only public pages that anyone can view without signing in. The lane Proxycurl got caught in is one CrustAPI is not in.
Where each Proxycurl endpoint maps
Most of what people used Proxycurl for has a direct home in CrustAPI's one LinkedIn endpoint, switched by a type= parameter.
| Proxycurl endpoint | CrustAPI | Notes |
|---|---|---|
| Person Profile | type=profile | Pass a linkedin.com/in/ URL |
| Company Profile | type=company | Pass a linkedin.com/company/ URL |
| Person Search | type=search | Search by keywords; add enrich=true for full profiles |
| Employee Listing | type=company&employees=true | Returns an employee sample |
| Jobs | type=jobs | Search by keywords and location, or one job by URL |
| Email / contact lookup | Not in the endpoint today | CrustAPI returns the public profile, company, posts, jobs, and people data. Email lookup is not part of it today. |
Proxycurl also had a Posts feed. CrustAPI serves recent public posts for a profile or company with type=posts. Migration questions? Email support@crustapi.com.
Proxycurl vs CrustAPI
| CrustAPI | Proxycurl | |
|---|---|---|
| Status in 2026 | Live | Shut down July 4, 2025 |
| Paid credits expire | Never | Yes, monthly on plans; PAYG expired too |
| Method | Public, logged out, no accounts | LinkedIn alleged fake accounts |
| Free tier | 3,000 / month, no card | Not available (closed) |
| Profile speed | ~0.8s server time | n/a |
| Email / contact lookup | Not offered | Had one (3-credit endpoint) |
Proxycurl figures from its shutdown post and its archived pricing page. CrustAPI figures from live calls on July 15, 2026. Something off? Email support@crustapi.com and we'll fix it.
One call, one profile
Here is the exact request and a trimmed version of what came back for a public profile on July 15, 2026. It returned HTTP 200 in 796ms of server time.
We drop LinkedIn's fake guest connections value of 8, so connectionsCount reads 500 the way a signed-out visitor sees it. Full field list is in the docs.
Why the credit clock matters
Proxycurl credits ran out
On subscriptions they reset each month, and even the pay-as-you-go credits expired after long inactivity. Buy a batch, pause a project, come back to a smaller balance.
CrustAPI credits don't
What you buy stays yours until you use it, and our terms say so in writing. A seasonal or on-and-off workload keeps its balance between runs.
You pay per successful call
One credit per successful call. A private or empty profile returns a clean 200 with a note and costs nothing, so failed pulls don't quietly bill you.
Nothing to get banned
There is no LinkedIn account behind these calls, so there is no login to restrict and no daily cap tied to a session. It reads the same pages a logged-out visitor sees.
Pricing, plainly
One credit is one result. The entry pack is $49 for 25,000 credits, which is $1.96 per 1,000 profiles. At volume it drops to $0.65 per 1,000. Every credit stays valid until you spend it.
$49 for 25,000. Good for testing and small runs. The 100k pack is $149, which is $1.49 per 1,000.
$6,500 for 10M. Steps in between: 500k $549 ($1.10/1k), 2.5M $1,999 ($0.80/1k). Nothing expires.
Free tier: 3,000 credits every month, no card. It resets each month and does not stack. Enriched people search is the one exception to the one-credit rule, billing 1 credit per full profile it returns.
Public, logged-out, no accounts
CrustAPI reads public pages while signed out. No account, no login, no fake profiles, no CAPTCHA solving. It only reads what any signed-out visitor sees. That is the lane courts have treated most favorably.
In hiQ Labs v. LinkedIn, the Ninth Circuit held in 2019, reaffirmed on April 18, 2022, that scraping data from a public website is not access "without authorization" under the Computer Fraud and Abuse Act, because a public page has no login gate to circumvent. To be fair about how that case ended: hiQ ultimately lost on contract, not hacking. A court found it had breached LinkedIn's User Agreement, which binds people who signed in and agreed to it, including by using fake accounts, and it settled for $500,000 and shut down. The point that stands is narrow: reading a public page is not hacking. The loss was for violating a contract hiQ had agreed to.
Two more recent rulings point the same way on public data. In Meta v. Bright Data (January 23, 2024, Judge Chen), a federal court ruled that the Facebook and Instagram terms do not bar logged-off scraping of public data. In X Corp. v. Bright Data (May 9, 2024, Judge Alsup), the court dismissed X's claims on a motion to dismiss, finding attempts to block copying of public data preempted by federal copyright law.
Those two decisions are district-court rulings, persuasive but not nationwide binding precedent, and they turned on those platforms' specific terms. So the honest summary is that courts have sided with public-data scrapers, not that scraping is settled law. And the contrast with Proxycurl is the whole point: what drew a permanent injunction there was fake accounts, which is logging in, not reading a public page. This is background, not legal advice. Check your own use with a lawyer.
Who this isn't for
A clean migration means being straight about the gaps. Skip CrustAPI if:
Switch in three steps
- Sign up. 3,000 free credits a month, no card, so the trial costs nothing.
- Swap the base URL. Point your profile and company pulls at crustapi.com/v1/linkedin?type=profile or type=company, with your key in x-api-key, and pass the LinkedIn URL as url.
- Map the rest: Person Search becomes type=search, Employee Listing becomes type=company&employees=true, Jobs becomes type=jobs. The docs list every field.
Common questions
Clay
Give your Proxycurl code a home
Sign up, run your real profile and company pulls, and see the JSON for yourself. It costs nothing on the free tier, and the credits you buy later never expire.
Get 3,000 free credits