LinkedIn Profile Data API in 2026: What You Can Fetch, and Store
The obvious place to get a LinkedIn profile is LinkedIn, and it is the one source that will only ever return your own.
No official LinkedIn API returns another member’s profile. LinkedIn’s profile endpoints read the authenticated member and nobody else, and its platform terms cap storage of most member profile data at 24 hours. Looking up an arbitrary profile means one of three third-party routes: a pre-built dataset, a scraper, or an account-based API acting on an account you connect yourself.
Freshness and storage rights are the obvious things to compare. Whether the provider is still trading in a year stopped being hypothetical in 2025, and that is now part of the same decision.
On this page
What LinkedIn’s own API returns about a member
The official profile surface reads one person: whoever authorised your app. Sign In with LinkedIn using OpenID Connect grants the openid, profile and email scopes, and /v2/userinfo returns a short, fixed record. There is no endpoint that takes a profile URL and returns that person’s data.
| Field group | What it holds | Access |
|---|---|---|
| OpenID Connect userinfo | sub, name, given_name, family_name, picture, locale, plus email and email_verified where granted |
Self-serve |
| Basic Profile | Adds maidenName, headline, vanityName and localised variants. Still no work history |
Partner approval |
| Position, Education, Certification, Course | The career fields most enrichment products are actually after | Partner approval |
| Skill, Language, Publication, Patent, Honor | The long tail of profile detail | Partner approval |
| Phone Number, Contact Interest, Website | Contact fields | Partner approval |
| Any other member’s profile | Passing a profile URL and getting that member’s record | No such endpoint |
| Matched profiles, partner surfaces | Profile data returned against CRM records a partner already holds (Sales Navigator Sync Services) | Partner approval |
Tutorials written before 1 August 2023 describe an API that no longer answers: the legacy Sign In with LinkedIn product, and with it r_liteprofile and r_emailaddress, was deprecated on that date. The partner-gated field groups above come from a reference last touched in 2023, so they describe documentation state and not necessarily what a partner receives today.
The storage rules that decide what you can build
Even with partner access, LinkedIn’s restricted-use-cases policy rules out the way most enrichment products work. The policy page was last updated on 29 August 2025 and states the limits directly:
“Member social activity data can only be stored for 48 hours, and most member profile data can only be stored for 24 hours.”
“Member data can’t be combined with your data, any other LinkedIn data, or any third party data to create, supplement, verify, or append to user profiles, leads, reference tables, or portions thereof.”
A third clause closes the remaining door. Member data may not be used “for advertising, sales, or recruiting use cases”, and the policy spells out what that covers: identifying prospects, creating leads, enhancing customer data in a CRM, building an audience list, ad targeting, account-based marketing, sending mass messages. A fourth forbids exporting member data from your application at all, including to your own customers.
Read together, those clauses describe a display integration with a 24-hour memory. Teams that discover that after building rebuild on a different route.
Three ways to get LinkedIn profile data
They differ on freshness, on whether they can act as well as read, and on what happens when LinkedIn changes its position.
| Route | What you get | Freshness | Main constraint |
|---|---|---|---|
| Dataset providers People Data Labs, Coresignal, Bright Data |
A pre-collected database, queried by identifier or filters. Bright Data publishes 905.4M+ LinkedIn records across profiles, companies, jobs and posts, from $250 per 100K records | As fresh as the last crawl | Coverage is broad and per-record accuracy is the vendor’s own claim |
| Scrapers Apify actors, Bright Data, self-hosted |
Whatever renders on a public page, parsed | Live at fetch time | Breaks on markup changes, reads only what is public, no authenticated actions |
| Account-based APIs gtm-api.com, Unipile, Linked API |
Live data through a LinkedIn account you connect, plus the ability to act on it | Live on first fetch, then cached (24h profiles, 7d detail, 7-30d company) | Throughput is bounded by what one account can safely do |
| Official API | The authenticated member only | Live | 24-hour storage cap, no export, no sales use |
A dataset genuinely wins on volume. If the job is to score ten million companies once, buy a dataset. Account-based access wins where the record has to be current at the moment you read it, or where reading is only half the job and something has to be sent afterwards. Vendor-by-vendor comparison of the whole category sits on Best LinkedIn API.
What the Proxycurl shutdown changed
One of the most-used LinkedIn profile APIs of the previous generation stopped trading in 2025, and the way it ended is now part of every buying decision in this category. LinkedIn filed suit against the operator, Nubela Pte. Ltd., on 24 January 2025 in the Northern District of California, case 3:25-cv-00828. Its founder, writing about the case in April 2026, counted six claims in the complaint: breach of contract, fraud and deceit under the California Civil Code, the Computer Fraud and Abuse Act, California’s Unfair Competition Law, trademark dilution under the Lanham Act, and misappropriation.
Proxycurl announced its shutdown on 4 July 2025. The founder’s stated reason was the cost of defending the case rather than a view on the merits, citing the American rule on legal fees against an opponent with, in his words, “more or less an unlimited war chest”. In April 2026 he confirmed the matter had settled and declined to discuss terms.
One nuance never got tested in court. The founder’s own account says the complaint quoted his earlier remark that fake accounts used to reach otherwise inaccessible data are obvious trouble, which is why providers that stay on logged-out public pages argued their position was different. Ask a profile-data vendor where the data comes from, whose account makes the request, and what happens to your pipeline if that vendor receives a letter.
Profile data through an account-based API
An account-based API reads LinkedIn as an account rather than as an app, which is why it returns what that account can see. GTM API reports 20,000+ connected accounts at under a 1% monthly ban rate on the stack that keeps them inside their limits. Its enrichment surface is a set of single-target calls, each returning a typed record, with results cached per target and method.
- Person profiles at three depths. A lite profile resolves identity, a basic profile adds the public summary, and a full profile returns headline, position, company, location, about, connection and follower counts, languages, education, and the heads of experience, skills and posts. Profile responses are cached for 24 hours.
- Career and activity detail as separate calls. Experience, skills, education, posts, certifications, interests, recommendations, and comment or reaction activity each resolve on their own, cached for 7 days.
- Contact details, scoped to a real relationship. The contact-info call routes to an account of yours that already holds a first-degree connection to the target, and returns whatever that person chose to expose to their connections, which is often nothing. With no such edge, the call is refused, and never routed to an unrelated account.
- Companies alongside people. Company profile, lite profile, public identifier and posts, cached from 7 to 30 days depending on how often the field changes.
- One target per call. There is no bulk enrichment endpoint. Batching is a client-side loop with an idempotency key on each request, and every response carries a ledger row recording which account executed it.
Where a call executes decides the ceiling. Route it to an account you connect and it spends that account’s daily read budget, which is what keeps the account safe and what caps your volume. Route it to the managed pool instead and there is no account to connect and no per-account limit, which is how larger enrichment jobs run. The per-account budgets are on the LinkedIn limits page.
Which profile-data route fits which build
Four shapes cover most of what teams actually ask for, and the split is not really about data quality. It is about whether the record has to be true at the moment you read it, and whether anything gets sent afterwards.
- Static scoring at scale. Ten million records, accuracy tolerant, no actions. Buy a dataset.
- One-off research. A few thousand public pages, throwaway output. A scraper is cheap and the fragility does not matter.
- A product that reads and then acts. Enrich a lead, then send an invite or a message from the same contract. Account-based, because the other two routes stop at reading.
- An integration displayed inside your own app. If you can live with a 24-hour memory and no export, the official partner route is the sanctioned one.
Reading and acting from one contract means inheriting the account ceiling along with the capability, so account safety belongs in the first version of that build.
# full reference: docs.gtm-api.com curl -X POST https://app.gtm-api.com/linkedin/v4/api/linkedin-enrichment/person-full-profile \ -H "Authorization: Bearer $GTM_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "linkedin_account_sid": "ln_ac_...", "nickname": "williamhgates" }'
Frequently Asked Questions
Is there a LinkedIn API to look up any member’s profile?
Not an official one. LinkedIn’s profile endpoints operate on the authenticated member only, and there is no endpoint that accepts another member’s URL or identifier. Looking up an arbitrary profile means a dataset provider, a scraper, or an account-based API that reads as a LinkedIn account you connect yourself.
How long can I store LinkedIn profile data?
Under LinkedIn’s restricted-use-cases policy, most member profile data may be stored for 24 hours and member social activity data for 48 hours. The same policy forbids exporting member data from your application and forbids combining it with other data to build or append to profiles and leads. Those limits ride with the member data itself, so partner approval widens the fields and leaves the rules in place.
What happened to Proxycurl?
LinkedIn sued its operator, Nubela Pte. Ltd., on 24 January 2025 in the Northern District of California, case 3:25-cv-00828, on six claims including breach of contract, the Computer Fraud and Abuse Act and trademark dilution. Proxycurl announced its shutdown on 4 July 2025, citing the cost of defending the case. The founder confirmed in April 2026 that the matter had settled, without disclosing terms.
Can I get someone’s email address from a LinkedIn profile?
Not from LinkedIn’s official API, which exposes the authenticated member’s own email and no one else’s. LinkedIn shows a member’s contact details to their first-degree connections in the interface, so an account-based API can return them where that relationship already exists. Dataset providers sell contact records built from other sources, and their accuracy is the vendor’s own claim.
What is the difference between a profile dataset and a profile API?
A dataset is a stored copy queried on demand, so it answers instantly at any volume and is exactly as current as the last crawl. A profile API resolves the record at request time, so it is current and bounded by whatever rate the source tolerates. Volume work suits a dataset. Anything where a stale field causes a wrong action suits an API.
Sources & Further Reading
- LinkedIn / Microsoft Learn, Sign In with LinkedIn using OpenID Connect (the
/v2/userinforesponse fields) and the legacy product, deprecated 1 August 2023 - LinkedIn / Microsoft Learn, Profile API reference (partner-gated field groups) and Basic Profile fields, both last updated 8 May 2023
- LinkedIn / Microsoft Learn, Restricted use cases (the 24-hour and 48-hour storage limits, the export ban and the sales-use ban), last updated 29 August 2025
- Steven Goh (Nubela / Proxycurl), “Goodbye Proxycurl”, 4 July 2025, and his April 2026 account of the case, in which he confirms the settlement
- Case reference: LinkedIn Corporation v. Nubela Pte. Ltd. et al, N.D. Cal. 3:25-cv-00828, filed 24 January 2025. The six claims are enumerated in the founder’s April 2026 post above. The public docket listing carries the case number and filing date only
- Vendor claims cited as their own: Bright Data LinkedIn datasets (905.4M+ records, from $250 per 100K, page updated 29 June 2026) · People Data Labs person enrichment · Coresignal employee data API
- GTM API, developer documentation
- Related: LinkedIn API: the developer guide · Sales Navigator API · How to get LinkedIn API access · LinkedIn API pricing · LinkedIn limits
Enrich a profile, then act on it, from one contract.
One typed contract covers the lookup and the message that follows it, on an account you own and inside the budget that keeps it alive. GTM API reports 20,000+ accounts at under a 1% monthly ban rate. Free plan, then from $10 per account at volume.
Last updated: August 2026 · Official-API fields, storage limits and case details verified against primary sources on 14 August 2026