Live Fetch

Live Fetch returns a LinkedIn profile or company as it stands right now, rather than the version already in our index. Use it when the answer has to be current: verifying a lead before an email goes out, hydrating a record when a user opens it, or re-checking a company before a deal review.

Two single-record operations do most of the work, with a batch pair for volume:

Identifiers we accept

You don't have to normalize your input first. Both endpoints accept several identifier formats.

For a person

Pass the person as identifier:

One field, four formats, no type discriminator to set — so a CRM column mixing profile URLs and bare slugs can go straight in.

For a company

Companies take a type and a value, so you tell us which kind you're sending:

The value field also accepts a Fiber company ID (comp_1441) if you're round-tripping an ID from an earlier response.

Asking for more depth

Both endpoints take optional flags that widen the response. They cost nothing extra — they just take longer, so turn them on only when you'll use the data.

getHistoricalHeadcount gives you company-wide totals through time. If you want the breakdown by department rather than the total, use the Depth Chart endpoints instead.

When a record isn't there

Profiles answer with a found discriminator, so you branch on one field rather than parsing an error. Companies return 404 when nothing matches.

Either way, check isCached404: it means the record is no longer reachable on LinkedIn, but we still hold data for it, and that data is what you got back.

When live fetch is unavailable

If a live fetch can't complete right now, we fall back to our cached copy when we have one. You get a 200 with the record, and advice explains what happened:

"advice": [
  "Live fetch is currently unavailable; this response is served from cached profile data.",
  "Retry later to get the freshest real-time LinkedIn data for this profile."
]

This fallback isn't charged. If we have no cached copy, the call returns 500 instead — retry later. isCached404 stays false here: it only marks a confirmed LinkedIn 404 (see above), so branch on advice.

Batches of up to 10,000

For volume, start a job and poll it, or subscribe to the batch_live_enrich.completed webhook to hear when it's done (see Webhooks).

1. Start the batch: send type (PROFILE or COMPANY) and up to 10,000 identifiers, in any of the formats above. You get back a taskId.
2. Poll for results: pass the taskId and page through results, up to 100 per page (take), following nextCursor until it comes back null.

One type per batch: a profile job takes profile identifiers, a company job takes company identifiers.

Your input is reconciled up front. The start response accounts for every identifier you sent:

numIdentifiersEnqueued + numDuplicatesSkipped + numMalformed = your input length

Duplicates are dropped before enrichment, so submitting the same profile twice doesn't charge you twice.

Each polled result carries its own status:

The progress object counts every state (pending included) against total, so you can render progress without a separate query. The job's own status moves PENDING → PROCESSING → COMPLETED or FAILED, and completedAt is set when it lands.

Using it effectively

• Use Live Fetch when the answer has to be current; use search or Kitchen Sink when indexed data will do. Live Fetch costs more per record.
• Leave the depth flags off unless you read the fields they add. They slow the call.
• Single-call rate limit is 600 requests/minute; allow around 30 seconds per call before timing out.
• Above a few hundred identifiers, use the batch endpoints rather than fanning out single calls: you get the duplicate and malformed accounting with them.
• Handle MALFORMED separately from NOT_FOUND. The first is a bug in your input; the second is a fact about the world.

Use cases

• Pre-send verification: confirm someone still holds the role your sequence assumes before the email goes out.
• CRM hydration: turn a column of mixed LinkedIn URLs and slugs into full records in one batch.
• Deal review: pull a company with getHistoricalHeadcount to see whether growth held up since you last looked.
• On-demand profile views: fetch when a user opens a record, so what they're looking at is what's live.

Cost

2 credits per profile or company fetched, on both the single and batch endpoints.

A record we determine isn't on LinkedIn counts as a fetch: we did the work, and you often still get the last known version back (that's what isCached404 is telling you). If a live fetch can't complete on our side, the fallback isn't charged: when we have a cached copy you get it back with advice to retry, and no credit is spent.

The chargeInfo on each response is what you were actually charged. See Billing & credits. Credit amounts are defaults — custom pricing is available for bulk use cases.

Related

• Contact Reveal — emails and phone numbers, which Live Fetch doesn't return.
• Kitchen Sink — lookups by email, name plus domain, and other messier identifiers.
• Search — find records when you don't already have an identifier.
• Tracker — get told when something changes, instead of re-fetching to check.