TL;DR
- HarvestAPI is a useful first evaluation for a team starting with known LinkedIn profile URLs and wanting a dedicated profile endpoint. Confirm the input, returned fields, freshness and current commercial terms for that endpoint.
- Crustdata fits a product that wants person and company data through an API and can distinguish indexed enrichment from more expensive live retrieval. A returned profile is not automatically a current employment verification.
- Apify provides a marketplace of separately maintained Actors rather than one universal LinkedIn API. The selected Actor's pricing, behaviour, maintenance and platform charges are part of the purchase.
- Bright Data offers both a scraper API and pre-collected datasets. A real-time request and a bulk snapshot answer different freshness and minimum-purchase questions.
- Coresignal is worth a trial when structured employee records or recurring datasets matter more than retrieving a specific page immediately. Match the record schema and refresh schedule to the application's needs.
- Buy a usable record, not a returned row. Check the identity, current role, source and observation time, then retain an explicit unknown when the evidence is incomplete. Public accessibility is not a blanket permission to collect or reuse personal data; assess your rights, applicable requirements and platform terms for the use case.
What are you asking the API to return?
A team can begin with a list of known public profile URLs, search for people matching a role, enrich a CRM contact, or monitor whether an existing contact changes jobs. Those are different data operations. A low price for returning a known profile says little about the cost of searching for that person, obtaining a work email or refreshing the record every month.
This guide focuses on research from known profile identifiers. The output needed for a B2B workflow might be a name, current role, current employer, location and a source reference. A profile can return successfully while still missing the current role, naming a parent rather than the operating company, or reflecting an earlier job. Those are buyer-relevant limitations, not merely incomplete JSON.
The LinkedIn-to-email API guide addresses a separate task: finding contact details from a profile. The LinkedIn outreach API comparison deals with connected sending accounts and webhooks. A data retrieval API is not, by itself, either an email finder or an outreach platform.
The options below use different source and delivery models. Confirm the collection route, rights and data provenance with each provider, and review what your own organisation is permitted to do with the information. We did not access private accounts, submit profile URLs to vendors, or conduct a comparative accuracy test for this article.
Five routes compared
| Option | Primary route | Starting point | Billing or operating unit | Best fit |
|---|---|---|---|---|
| HarvestAPI | Dedicated profile and adjacent endpoints | Known profile ID or URL | Endpoint and returned-result terms; obtain current pricing | A focused profile API integration |
| Crustdata | Person search and enrichment | Profile URL or other identifiers | Credits by result and requested fields | An API product combining people and company research |
| Apify | A selected third-party Actor on a platform | Usually a profile URL or a batch, depending on Actor | Actor events plus applicable platform usage | A team wanting a specific reusable extraction workflow |
| Bright Data | Profile scraper API or pre-collected data | Known profile request or a dataset order | API records versus dataset minimum and refresh terms | A team comparing on-demand and bulk delivery |
| Coresignal | Structured employee API and datasets | Employee identifier and supported search criteria | Record and field credits or dataset contract | A pipeline needing structured employee records at scale |
Account requirements are endpoint-specific. A vendor may offer public-profile retrieval and also offer separate features that use a connected account. Do not provide a user's session cookie to an unrelated service merely because its profile endpoint exists. Check the selected route and use the minimum authorised access for the job.
1. HarvestAPI: a dedicated profile endpoint
HarvestAPI's profile documentation describes a profile endpoint, with separate endpoints for other LinkedIn entities and operations. It is a sensible candidate when a developer wants a narrowly scoped API call rather than maintaining a generic browser workflow. Start by checking the accepted identifier, request limits and response fields for the exact endpoint.
The documented query accepts a profile URL, public identifier or profile ID. Its main=true option returns a shorter version, including at most five experience entries, and the vendor says it uses fewer credits. An optional findEmail request costs more credits and can return catch-all addresses marked risky. Choose the shorter response only if its coverage is sufficient for the use case, and treat a risky email as an uncertain status rather than a verified deliverable address. Confirm the current credit rates for the selected options.
For a known URL, the useful test is not simply whether the response includes a name. Check that the profile belongs to the intended person, that employment entries have a clear current-versus-former interpretation and that missing data is distinguishable from an empty field. If the downstream system requires a timestamp, establish whether it means retrieval time or an observation time supplied with the underlying data.
HarvestAPI also documents a custom accounts guide for other workflows. Do not generalise those account-dependent features to the public-profile route, or assume every endpoint has the same access requirements. Confirm the provider's current source, access method and permitted use in the order form and documentation.
We could not establish a current public dollar tariff for the exact profile endpoint from the pricing page accessed for this guide. Obtain the plan, rate, minimum, retry and failed-result treatment before budgeting. An unavailable pricing page is not evidence that the vendor is quote-only.
2. Crustdata: indexed person enrichment
Crustdata's People API accepts person identifiers, including a profile URL, for enrichment and separately supports search. Its published pricing on that page lists cached person enrichment from one credit per profile, with contact data and live retrieval priced higher. Confirm current field charges and endpoint eligibility before purchasing.
This route is useful when a product also needs company or person attributes beyond what a single profile page returns. It can accept a known identifier, but its normalised record should not be represented as a verbatim copy of a LinkedIn page unless the provider documents that provenance. Record which fields are requested, the returned entity ID and whatever source or freshness information the response actually supplies.
A key pricing distinction: one credit is not automatically one US dollar. The job's cost depends on the credit pack or contract and whether a work email, phone, live request or other premium field is added. Ask for an estimate covering 10,000 requests, expected matched records and the specific field set.
Crustdata is less compelling when the only requirement is a one-off, low-volume page extraction and the application does not use the broader record. Its search endpoint also represents a different job from enriching the known URLs used in the model below.
3. Apify: choose an Actor and operating model
Apify is a platform where individual developers publish Actors. The relevant purchase is the specific Actor plus the Apify platform, not “Apify” as if every scraper had the same code, support or billing. Check the maintainer, recent changes, input schema, output fields and documented error handling.
For example, the Data Slayer profile Actor lists a price from $4 per 1,000 profile results, accepts profile URLs and describes structured JSON/CSV output. The listing says it does not need LinkedIn login cookies for that workflow. Its pricing tab also lists an $8-per-1,000 profile enrichment flag event and says platform usage is included in these event prices. The Starter subscription is shown at $19 per month with $19 to spend on Actors. Treat these as this Actor's documented terms, not terms for every Actor; check whether the flag event applies to the fields you need and how your plan's included usage is allocated.
The marketplace model provides choice, but transfers selection and maintenance diligence to the buyer. A run may be billed by a result event, start event, compute, subscription or another combination. Failed inputs, partial profiles and retried runs may be treated differently across Actors. Retain the Actor version and the input/output contract used in your test.
A community-maintained Actor can be a useful first evaluation for a bounded extraction job. A team with strict support or data-processing requirements should establish who operates the Actor, what data leaves the account, and how provider changes are handled before putting it in a production pipeline. No Actor execution is claimed here.
4. Bright Data: on-demand retrieval or datasets
Bright Data's LinkedIn scraper product and its LinkedIn datasets are separate buying routes. The API is relevant when the application needs to request particular records; the dataset is relevant when a team can work with a prepared collection and its delivery schedule.
The dataset page describes profile, company, job and post datasets, with formats and refresh options varying by purchase. Its published example starts at $250 for 100,000 records for a one-time dataset order. That is not the price of making 10,000 fresh profile API requests. A snapshot can be economical at scale while still being unsuitable for a time-sensitive “is this person still in this role?” decision.
Ask how the selected dataset was collected, the actual collection window, the fields, licensing and update cadence. If the application needs a fresh check on a subset of accounts, compare a hybrid architecture: retain permitted dataset information and request a current observation only where the business rule requires it.
Bright Data is a stronger candidate when the team values multiple delivery formats, procurement options or a mix of API and datasets. It is less directly comparable to an individual low-cost profile Actor unless both are scoped to the same output and refresh requirement.
5. Coresignal: structured employee records
Coresignal offers employee and company records through APIs and datasets. For an application building a persistent account or talent data pipeline, a structured employee record and a repeatable delivery contract may matter more than retrieving a particular webpage at request time.
Its pricing page lists a full employee record at 20 credits, a core record at 10 credits, and separate field bundles. Employee contact enrichment is listed separately, including ten credits per email found. The public page lists a Mini plan at $49 per month with 2,500 credits and larger plans with different allowances. Confirm plan and endpoint eligibility and the price for the specific fields before treating this as a per-profile quote.
For known profile URLs, validate how the provider matches an input to an employee record, what a missing record returns and whether the record's freshness is sufficient. A profile URL is an identity clue, not proof that two providers have resolved it to the same person or current employer.
Coresignal is a useful shortlist choice for structured records or datasets when the buyer can use the wider schema. It is not necessarily the simplest purchase for a one-off handful of profile reads, and its employee-record credit should not be compared directly with a returned-page fee without matching the output.
What does a 10,000-profile job cost?
Suppose a team already holds 10,000 profile URLs and needs current role and employer for account research. In an illustrative batch, 8,000 inputs return a profile and 6,000 records pass the identity, current-role and company acceptance rule. Those figures are assumptions, not measured vendor results.
| Route | Illustrative chargeable unit | What the model can say | What remains outside it |
|---|---|---|---|
| HarvestAPI | Exact endpoint and plan terms | Obtain a current quote for 10,000 inputs and retries | Result, miss, concurrency and support terms |
| Crustdata | From one credit per matched base person profile | 8,000 matched base profiles would use at least 8,000 credits under that assumed unit | Credit-to-dollar conversion, optional fields, search and live access |
| Selected Apify Actor | $4 per 1,000 profile events; flag event listed separately | At the base event rate, 8,000 results would incur $32 in profile-event charges | Flag-event eligibility, plan allowance and any other applicable charges; actual returned and accepted counts |
| Bright Data dataset | Published $250 minimum example for 100,000 records | A 100,000-record snapshot is a different purchase from 10,000 on-demand lookups | Freshness, filtering, API tariff and licensing |
| Coresignal full employee record | 20 credits per full record | 8,000 full records × 20 = 160,000 credits | Plan selection, additional fields, minimums, misses and permitted use |
The Apify number is arithmetic using one named Actor's published profile-event rate, assuming that rate applies to every billed result. If 6,000 of the 8,000 returned records were accepted, $32 ÷ 6,000 is about $0.0053 per accepted record for the profile events alone. The Actor's listed event rates include platform usage, but the model excludes any enrichment-flag events, other Actor usage and the work of checking results. Do not add the full $19 Starter subscription to the $32 without accounting for its included $19 allowance. This is not proof that the Actor would return 8,000 valid profiles; inspect the current plan and event pricing before budgeting.
For Coresignal, the illustrative 160,000 credits exceed the 150,000 included in the displayed $1,000 monthly Growth plan. The published comparison also lists a $1,500 monthly Premium plan with one million credits. Ask how the provider would bill the exact workload and whether a smaller overage or another arrangement is available; do not price 160,000 credits as though they fit in Growth's included allowance.
Refresh frequency changes the purchase. Three equal monthly refreshes at that hypothetical returned volume would produce $96 in Actor result charges before the same exclusions. If only 2,000 records need a monthly refresh, model that policy separately. The freshness requirement should come from the decision the data supports, not an arbitrary promise to refresh everything.
For the other vendors, build the same ledger with the actual quotation: attempted inputs, matched results, required field add-ons, retries, credits and dollar conversion, minimum commitment and permitted reuse. A dataset minimum should not be hidden by dividing $250 by only the records the team expects to use unless the other purchased records have a genuine purpose.
How do you test freshness and identity?
Construct a small, authorised fixture before buying volume. For example, use 30 profile identifiers the organisation is permitted to process: ten stable roles, ten recent job changes, five similar names and five missing or inaccessible profiles. This is a proposed evaluation design, not a test we ran.
For each input, record five outcomes:
- Identity: does the response refer to the intended person and identifier?
- Current role: is the returned role explicitly current, historical or ambiguous?
- Company: does the returned organisation match the operating entity needed by the CRM?
- Freshness: does the provider supply an underlying observation date, a retrieval date, both or neither?
- Availability: is a field absent, inaccessible, unreturned by that plan or actually empty?
Where the reference evidence is uncertain, do not score a provider wrong merely for disagreeing with an unverified CRM entry. Mark the case unresolved. A small fixture reveals integration and identity failures, but it cannot establish a universal coverage ranking across countries or industries.
If the person recently changed jobs, a profile that still shows the former role may be a stale observation, while a different provider may have matched a namesake. Those failure modes demand different handling. Keep the source record and reason for review rather than reducing both to a single “bad result” percentage.
What should the integration retain?
A useful response contract separates what the vendor returned from what the organisation has decided to do with it. For each accepted record, retain the internal person ID, input URL or authorised identifier, provider record ID, current-role candidate, company identifier, retrieval time and any available source-observation time. Include the version of the matching rule and a status such as accepted, ambiguous, unavailable or stale.
For example, a response received today can report a role observed months ago. retrieved_at identifies the API call; it is not the same as observed_at. If the vendor supplies no observation date, store that absence and decide whether the workflow needs additional verification. Do not manufacture a freshness date from the request timestamp.
Make retries idempotent. A temporary error should not create another CRM person, and an asynchronous job should be polled using its original job reference where the provider supports it. Reconcile costs as well as data: a repeated successful profile or platform run may be chargeable even if the application later discards it.
Use the GTM engineering playbook to define the account owner, review queue and permitted downstream actions. An enrichment result should not automatically start outreach or overwrite a customer-confirmed title without an explicit rule.
Which API should you choose?
Start with HarvestAPI for a focused profile endpoint and Crustdata when person and company enrichment belong in the same API-driven product. Compare their exact inputs, output fields, documented source access and prices on the same authorised fixture.
Evaluate a specific Apify Actor when its input, output and support model meet a bounded job. Evaluate Bright Data when the choice between on-demand requests and a bulk dataset is itself important. Consider Coresignal when structured employee records and a broader data pipeline justify its credit model.
None is the default winner for every geography, data right or refresh window. The correct route is the one that returns the fields your application can lawfully use, at the freshness and total cost the decision requires.
FAQ
Is a LinkedIn data API the same as a LinkedIn outreach API?
No. A profile data API retrieves or supplies information about a person. An outreach API may connect an account, manage conversations or send messages. Some vendors offer multiple products, but their access and billing can be different.
Does a successful profile response prove someone still works at a company?
No. It proves the API returned a response. Check the underlying observation date where available, distinguish current from historical roles and corroborate important changes before updating ownership or acting on them.
Can I use a low per-profile price for a 10,000-record budget?
Only after checking what creates a billable result, platform fees, failed attempts, rate limits and the fields included. Divide the complete cost by records that satisfy your acceptance rule, not merely by inputs submitted.
Is publicly visible information always available for any use?
No. Public visibility alone does not settle the terms, rights or obligations attached to collection, storage and downstream use. Review the provider's contract, platform terms and applicable requirements for your specific workflow.
Do I need a LinkedIn login for every route?
No single answer covers every endpoint. Some providers describe public-profile routes that do not require your account credentials; other functionality may use a connected account. Verify the selected endpoint's requirements before giving any service access.





