Migrate from Proxycurl to LinkFinder AI

Proxycurl shut down in July 2025. This is a developer's guide to moving your LinkedIn data pipeline to LinkFinder AI — with a full endpoint map, before/after code, and an architecture that doesn't depend on LinkedIn accounts.

Last updated: May 30, 2026

Get a free API key Read the API docs

What happened to Proxycurl

For years, Proxycurl was the LinkedIn data API that developers reached for first. It was clean, well-documented, fairly priced, and reliable — a genuinely good product built for engineers. Credit where it's due: it set the bar for what a LinkedIn enrichment API should feel like.

That ended in 2025. In January 2025, LinkedIn (owned by Microsoft) filed a federal lawsuit alleging Proxycurl operated large numbers of fake accounts to collect member data. Facing the cost of fighting a multi-billion-dollar opponent, Proxycurl settled and shut down in July 2025. The API endpoints went dark, the documentation was pulled, and under the settlement collected LinkedIn data had to be deleted.

If you're reading this, your pipeline is probably broken. Proxycurl no longer responds, so the job now is to repoint your integration at a provider that does the same work. This guide maps every Proxycurl endpoint to its LinkFinder AI equivalent and shows the code changes.

One thing worth being deliberate about while you migrate: the reason Proxycurl was vulnerable was its reliance on LinkedIn accounts. LinkFinder AI's B2B Data Lookup retrieves business-contact and company data without using your own LinkedIn account, which removes both the ban risk on your side and the specific exposure that ended Proxycurl. No provider can promise it will never face litigation, but the underlying approach is different.

Proxycurl vs LinkFinder AI at a glance

An honest side-by-side. Proxycurl genuinely did some things well — this isn't a clean sweep.

Criteria
LinkFinder AI
Active
Proxycurl
Shut down
Status (2026)
Live & supported
Offline since July 2025
Uses LinkedIn accounts
No — zero ban risk
Yes — basis of the lawsuit
API surface
1 endpoint, type field
Many separate endpoints
Developer docs & DX
(was excellent)
Natural-language lead search
(filters only)
Bulk profiles per call
Up to 100
1 per call
Pricing model
1 credit = 1 request
Credit / subscription
Free credits to start
100 free
Currently accepting customers

Endpoint mapping: Proxycurl → LinkFinder AI

Proxycurl split work across many endpoints. LinkFinder AI uses a single endpoint — POST https://api.linkfinderai.com — where the type field selects what you want. Here's how the old endpoints map across.

Proxycurl endpointWhat it didLinkFinder AI equivalent
Person ProfileProfile data from a LinkedIn URLB2B Data Lookup
Person SearchFind people by filters (title, geo, company)AI Lead Finder · leads_finder_ai
Person LookupResolve a person from name + companyAI Lead Finder · leads_finder_ai
Role LookupFind the decision-maker at a companyAI Lead Finder · leads_finder_ai
Company ProfileCompany data from a company URLCompany Enrichment
Company LookupResolve a company from its nameCompany Enrichment · company_name_to_website
Employee CountCompany headcountCompany Enrichment
Contact API (email / phone)Work email & phoneB2B Data Lookup
One thing to confirm: the request type values shown in code below that aren't in your published docs (e.g. profile / company lookups) are placeholders — check the exact strings in your dashboard under API Documentation. The documented types are leads_finder_ai and company_name_to_website.

Before & after: the code change

Looking up a person (Proxycurl Person Profile → B2B Data Lookup)

Before — Proxycurl (offline)
# No longer responds since July 2025 curl -X GET \ 'https://nubela.co/proxycurl/api/v2/linkedin?url=LINKEDIN_URL' \ -H 'Authorization: Bearer PROXYCURL_KEY'
After — LinkFinder AI
curl -X POST "https://api.linkfinderai.com" \ -H "Content-Type: application/json" \ -H "Authorization: Bearer YOUR_API_KEY" \ -d '{ "type": "profile_lookup", "input_data": "LINKEDIN_URL" }'

Finding leads (Proxycurl Person Search → AI Lead Finder)

This is an upgrade, not just a swap. Instead of building filter parameters, describe the audience in plain English and get up to 100 profiles per call.

# Python — replaces Person Search + Person Lookup + Role Lookup import requests resp = requests.post( "https://api.linkfinderai.com", headers={"Authorization": "Bearer YOUR_API_KEY"}, json={ "type": "leads_finder_ai", "input_data": "VP Sales at B2B SaaS startup in the United States", "fetch_count": 50 }, ) data = resp.json() if data["status"] == "success": leads = data["result"]

Company data (Proxycurl Company Profile / Lookup → Company Enrichment)

// Node.js — resolve a company from its name const res = await fetch("https://api.linkfinderai.com", { method: "POST", headers: { "Content-Type": "application/json", "Authorization": "Bearer YOUR_API_KEY", }, body: JSON.stringify({ type: "company_name_to_website", input_data: "Tesla", }), }); const { status, result } = await res.json();
Response shape changed too. Proxycurl returned the object directly. LinkFinder wraps every response in a status field ("success" or "error") and a result field. Update your parsing to read result, and note that a successful call can still return result: null if no data was found (it still costs 1 credit).

Migrate in four steps

Get your LinkFinder API key

Sign up for a free trial — your key is in the dashboard under Settings → API Key. You start with 100 free credits, enough to test your whole migration before paying.

Swap the base URL and auth header

Point requests at https://api.linkfinderai.com and replace your Proxycurl key with Authorization: Bearer YOUR_API_KEY. All requests are POST with a JSON body.

Map each old endpoint to a request type

Use the mapping table above. Most Proxycurl person/search calls collapse into a single leads_finder_ai call; company calls become Company Enrichment.

Update response parsing & retries

Read status and result instead of the raw object. Handle 429 with exponential backoff (1s → 2s → 4s) and treat result: null as "not found", not an error.

Why teams pick LinkFinder AI after Proxycurl

The same jobs, a simpler API, and an approach that removed the single biggest risk Proxycurl carried.

No LinkedIn accounts

B2B Data Lookup pulls business-contact and company data without using your LinkedIn account — no ban risk for you, and not the architecture that ended Proxycurl.

One endpoint, less code

A single POST endpoint with a type field replaces a dozen Proxycurl routes. Fewer integration points, less to maintain. See the API docs.

Lead search, upgraded

Describe your ideal customer in plain English and get up to 100 matching profiles per call — a step up from Proxycurl's filter-based Person Search.

Proxycurl migration FAQ

Is Proxycurl still available in 2026?
No. Proxycurl shut down in July 2025 after LinkedIn (owned by Microsoft) filed a federal lawsuit in January 2025. The API endpoints were taken offline and the documentation removed, so anyone who relied on Proxycurl needs a different LinkedIn data provider.
Why did Proxycurl shut down?
LinkedIn's January 2025 lawsuit alleged Proxycurl created hundreds of thousands of fake accounts to scrape member profiles. Rather than fight a prolonged case against Microsoft's resources, Proxycurl settled and wound down in July 2025. Under the settlement, collected LinkedIn data had to be deleted.
What happens to my Proxycurl integration?
Proxycurl's API no longer responds, so any pipeline that called it is broken. You repoint it at a new provider. With LinkFinder AI that means changing the base URL and auth header, mapping old endpoints to request type values, and reading the new status/result response shape. Most teams finish in an afternoon.
Is LinkFinder AI a drop-in replacement?
Not byte-for-byte. Proxycurl exposed many endpoints; LinkFinder uses one POST endpoint with a type field. Every job Proxycurl did — profile lookup, person search, company data, employee count — has an equivalent, so it's a remapping exercise, not a rewrite. See the API documentation.
Will LinkFinder AI get shut down like Proxycurl?
No provider can promise immunity from litigation. The factual difference: LinkedIn's complaint against Proxycurl centered on its use of fake LinkedIn accounts. LinkFinder AI's B2B Data Lookup does not use or require any LinkedIn account, so the specific architecture that ended Proxycurl isn't the one LinkFinder relies on.
How long does migration take?
Typically an afternoon: swap the base URL to https://api.linkfinderai.com, change the auth header to your LinkFinder key, map endpoints to request types, and update response parsing.
How does the pricing compare?
LinkFinder uses a simple model — 1 credit equals 1 request, regardless of endpoint — and new accounts start with 100 free credits. Check the pricing page for current plan limits.
Can I search for many leads at once, like Person Search?
Yes. The AI Lead Finder (leads_finder_ai) replaces Person Search, Person Lookup, and Role Lookup. Describe your audience in plain English and get up to 100 matching profiles per call, 1 credit each.
Which languages and SDKs are supported?
It's a standard REST API that accepts JSON, so any language that can make an HTTP request works — Python, Node.js, Go, Ruby, PHP, Java, and more. The docs include cURL, Python, and Node.js examples.

Ready to move off Proxycurl?

Get a free API key with 100 credits, point your pipeline at LinkFinder AI, and have your migration tested before lunch.

100 free credits • No LinkedIn account required • REST + JSON