Guide
LinkedIn API rate limits: how they work, and where to find yours
LinkedIn's official API caps how many calls your application, and each member using it, can make in a day. The counters reset at midnight UTC, and LinkedIn doesn't publish the standard numbers: you look them up for your own app. Here is how the limits work, from LinkedIn's API documentation, where to find yours, and how pacing works when you call LinkedIn through members' own accounts.
By Lempile · Updated
How LinkedIn's API limits work
Every call to LinkedIn's API counts against a daily limit, over a 24-hour day that resets at midnight UTC. Two kinds of limit apply at the same time:
- Application: the total calls your app can make in a day.
- Member: the calls one member's token can make through your app in a day.
Each endpoint has its own limit. A call over the limit gets a 429 response. LinkedIn notes that it may also answer 429 to protect its infrastructure, and that those clear up on their own.
Where to find your limits
LinkedIn doesn't publish its standard rate limits. To see yours, open your app in the Developer Portal and go to its Analytics tab. It lists usage and limits only for the endpoints your app has called at least once today (UTC), so make one test call to an endpoint first if you want to see its limit.
Developer Admins also get an email when the app goes past 75% of its quota. The alert comes about one to two hours late and only covers app-level limits, not member-level ones.
The limits LinkedIn does publish
| Limit | Value | Source |
|---|---|---|
| Reset | Every day at midnight UTC | Rate Limiting |
| Over the limit | HTTP 429 | Rate Limiting |
| Usage alert | Email at 75% of the app's quota, 1–2 hours late | Rate Limiting |
| Share on LinkedIn | 150 requests per member a day | Share on LinkedIn |
Handling 429s well
- Treat a 429 as "come back after the reset", not "retry now". A retry loop only spends the next day's quota faster.
- Count calls per member as well as per app, since either limit can stop you.
- Cache what you read, and spread calls over the day.
- Check the Analytics tab after adding an endpoint, so you know its limit before your users find it.
When you call LinkedIn through members' accounts
Most of what products want from LinkedIn, such as messages, invitations and other members' profiles, isn't in the official API (what's open and what isn't). Account-based APIs reach it through accounts your users connect themselves. There's no developer quota there, but LinkedIn still watches each account's pace, so the API has to pace it like a person.
Lempile paces every connected account on its own:
| Pace | Value | Over it |
|---|---|---|
| LinkedIn calls at once | One | A call waits its turn, up to 60 seconds |
| Gap between calls | 1.5 to 3 seconds | Calls wait; nothing is refused |
| Calls in any 24 hours | 1,500 | 429 with retry_after |
| Writes in any 24 hours | 150 | 429 with retry_after |
| After LinkedIn says “too many” | 15-minute rest | 429 with the seconds left |
Details are in Lempile's errors and limits docs, and LinkedIn's own limits on invitations are in LinkedIn connection and invitation limits.