LinkedIn API Rate Limits in 2026: The Numbers LinkedIn Does and Doesn’t Publish
You can design around a number you know. LinkedIn gives you one you can only find by hitting it.
LinkedIn throttles its API on two levels, per application and per member, both counted over a 24-hour day that resets at midnight UTC. It publishes no general figure and concrete numbers for only three products. Your real limits live in the Developer Portal, on the Analytics tab of your app, and only for endpoints it has already called that day.
That last detail is the one that costs a weekend. An endpoint you have never called shows no limit at all, so finding your quota means spending some of it. What you find is your own app’s number, and another app calling the same endpoint can see a different one.
On this page
The two-tier model and the reset clock
Every LinkedIn API call is counted twice, once against your app’s total and once against the token that made it, with the member count kept per app. LinkedIn’s rate-limit documentation states the window in one sentence:
“Rate limits specify the maximum number of API calls that can be made in a 24 hour period. These limits reset at midnight UTC every day.”
That is a calendar day, not a window that rolls forward from your first call. What the same page does not give you is a number: it describes the model, the reset, the Analytics tab and the meaning of a 429, and leaves every per-endpoint figure to the portal.
Every app on the platform gets its allowance back at the same instant. That does not consume your quota, because quotas are counted per app. It does mean that if LinkedIn is shedding load anywhere, just after the reset is where you are most likely to meet it, so spreading work across the day is safer than queuing it against midnight.
Where your actual numbers are visible
Open linkedin.com/developers/apps, select the app, and go to the Analytics tab. It lists the endpoints your app has called during the current UTC day with the quota attached to each. Endpoints you have not touched today are absent, so the only way to reveal an unlisted limit is to make one call and refresh the page.
A separate quota governs OAuth refresh tokens, and it gained its own monitoring in October 2025: percentage of quota used, total allowed per app per day, current throttling status and a countdown to the 00:00 UTC reset. Apps that mint a fresh token for every user session tend to meet this one first.
- Reveal before you rely. Call each endpoint your integration needs once, in a controlled window, then read the Analytics tab to record the real numbers.
- Cache what you are allowed to keep. Storage of member data is capped at 24 hours for most profile fields, so caching strategy and the platform terms have to be designed together.
- Keep development traffic on a separate app. Test volume otherwise consumes the quota your production users need, and both count against the same application-level bucket.
- Back off on the first 429. Retry storms burn the remaining allowance and make the recovery window longer.
The numbers LinkedIn does publish
Three products carry concrete figures in the documentation. Everything else is portal-only.
| Product | Tier | Published limit |
|---|---|---|
| Community Management | Development | 500 calls per app and 100 calls per member per 24 hours. Batch operations blocked, webhook push notifications disabled |
| Standard | No restrictions stated | |
| Verified on LinkedIn | Development and Lite | 5,000 calls per day |
| Plus | Custom | |
| Advertising | Development | Ad-account entitlements, counted in accounts: read unlimited accounts you administer, create 1 test account, edit up to 5 |
| Standard | No ad-account restrictions stated. Call quotas not published | |
| Everything else | Any | Not published |
Consumer products, including Sign In with LinkedIn and Share on LinkedIn, publish nothing. Figures circulating for them come from other developers’ portal readings, which are that app’s limits and not a platform constant.
A 429 does not always mean your quota
A 429 from LinkedIn carries two different meanings, and the response body does not say which one you have.
- Quota exhaustion. You have spent the application-level or member-level allowance for that endpoint on that UTC day.
- Platform load shedding. LinkedIn is protecting its own infrastructure. The refusal has nothing to do with your usage and it clears on its own.
A 429 that lands after steady traffic on one endpoint, close to a number you have seen on the Analytics tab, is a quota. A 429 that appears suddenly across endpoints you have barely used, on traffic that worked an hour earlier, is more likely platform-side, and best waited out before you re-architect anything. Log the endpoint, the UTC hour and the call count on every 429, because that record is what makes the two distinguishable later.
The alert, and its blind spot
Developer Admins receive an email when an app reaches 75% of an application-level quota, a feature that shipped with the September 2025 Marketing release. The alert arrives one to two hours after the threshold is crossed, so on a fast-moving endpoint the limit can be reached before the warning lands. And it fires for application-level breaches only, never member-level, so an integration where one heavy user exhausts their own quota gets no email at all. Multiple endpoints crossing the line in one day produce multiple alerts.
Why an account-based API hits a different wall
LinkedIn’s published throttles govern an app calling its platform. An account-based API acts as a LinkedIn account instead, so the scarce resource becomes the account itself, and LinkedIn watches how much a member does per day. LinkedIn’s Terms of Service do not permit third-party automation, so every tool in this category carries that risk, and what separates them is how hard they work to keep an account inside its own ceiling. GTM API reports 20,000+ connected accounts at under a 1% monthly ban rate on the model below.
GTM API’s own per-account budgets, self-reported from the engine that runs them, are held per connected account and per action type across 16 buckets, each with a daily budget and its own spacing between calls. Connection requests default to 35 a day at roughly ten minutes apart, messages to 50, InMail to 30, profile visits to 50. Those are ceilings for a fully warmed account, and a new one starts far below them and ramps as it builds history.
The recovery clocks, also self-reported, differ by action. When LinkedIn refuses an invitation the account pauses on invitations for about an hour, an InMail refusal pauses for about three, and other refusals hold until the account’s own local midnight. Counting is on success, so a failed call does not spend the budget. Our own throttle returns 429 with the reason and a retry hint inside the error body, and the full model with every bucket sits on the LinkedIn limits page.
Frequently Asked Questions
What are LinkedIn’s API rate limits?
LinkedIn applies an application-level limit covering all calls your app makes and a member-level limit covering calls made with one member’s token, both counted over a 24-hour day. It publishes no general figure, and limits vary per endpoint. Three products carry documented numbers: Community Management on the Development tier at 500 calls per app and 100 per member per day, Verified on LinkedIn at 5,000 calls a day on Development and Lite, and the Advertising API, whose Development tier is stated in ad-account entitlements.
When do LinkedIn API rate limits reset?
At midnight UTC, daily. The window is the calendar day, so every app on the platform regains its allowance at the same moment. Nobody else’s traffic touches your quota. If LinkedIn is shedding load, though, the minutes after the reset are where you are most likely to meet it.
Why am I getting a 429 from the LinkedIn API?
Either you have exceeded an application-level or member-level quota, or LinkedIn is shedding load to protect its infrastructure. The second kind recovers on its own and is unrelated to your usage. The response does not say which one you have hit, so log the endpoint, the UTC hour and your call count on every 429 and compare that against the Analytics tab.
Where can I see my LinkedIn API quota?
In the Developer Portal at linkedin.com/developers/apps, select your app and open the Analytics tab. It shows only the endpoints your app has called during the current UTC day, so an endpoint you have never called displays no limit. To find one, make a single call and refresh the page. Refresh-token quota is tracked separately on the same dashboard.
Do LinkedIn API rate limits apply per app or per user?
Both, at the same time. The application-level limit aggregates every call your app makes across all users. The member-level limit applies to each member’s token individually, within your app. A call can be refused because your app is out of allowance or because that one user is, and the quota alert email only covers the first case.
Sources & Further Reading
- LinkedIn / Microsoft Learn, API rate limits (the two tiers, the midnight UTC reset, the Analytics tab behaviour, 429 as both throttle and load shedding, and the 75% alert), last updated 20 August 2025
- LinkedIn / Microsoft Learn, Increasing API access (Community Management and Advertising Development-tier quotas), last updated 28 July 2026
- LinkedIn / Microsoft Learn, Verified on LinkedIn (the 5,000 calls per day figure), last updated 10 December 2025
- LinkedIn / Microsoft Learn, Developer portal tools (refresh-token quota monitoring, added October 2025) and recent Marketing changes (the 75% alert shipped in version 202509)
- LinkedIn / Microsoft Learn, Restricted use cases (the 24-hour storage cap that shapes any caching strategy), last updated 29 August 2025
- GTM API, developer documentation
- Related: LinkedIn limits · LinkedIn API: the developer guide · LinkedIn connection limits · Safe LinkedIn automation · LinkedIn API access
Outreach limits that hold the account inside its own ceiling.
Per-account budgets, randomized pacing, and a warm-up that raises capacity as an account earns it. 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 · Throttle model, published quotas and alert behaviour verified against LinkedIn’s developer documentation on 14 August 2026