Skip to main content
Smart Limits are the allowances Edges enforces on each LinkedIn action, per identity. We set them below what LinkedIn tolerates, so that an account working hard through Edges stays a working account.
These are our limits, not LinkedIn’s. LinkedIn’s own caps sit above them and are neither published nor stable — see How Limits Work for how the two relate.
What happens when a limit is reached? In live mode the request fails with a limit error. In async and schedule modes the work is postponed and resumes automatically once the window frees up.

How Smart Limits work

Example. If it’s Wednesday at 2 PM, Edges looks at the trailing 24 hours. If the quota for an action is spent, async and schedule work for that action is postponed; by 3 PM some of it has aged out and the work resumes. Identity A reaching its connection-request limit does not stop Identity A from extracting profiles, and does not affect Identity B at all.
Edges counts only the actions performed through the platform. Activity done by hand on LinkedIn, or through another tool, is invisible to us and still counts against the account with LinkedIn.

New identities start lower

A newly connected identity does not receive its full allowance immediately — it builds up over the first days. This is the most common reason a limit is reached earlier than the tables below suggest.

Identity Ramp-Up

How capacity grows, how to check where an identity currently sits, and how to lift it.

Quotas

Quotas by window

Each number is the cap for one rolling window, per identity. It applies to a LinkedIn account without a paid subscription. LinkedIn Premium and Sales Navigator are named in the cell when their cap differs. On a Sales Navigator limit, 0 is the cap for an account without Sales Navigator. InMail’s long window is 31 days, in the Per 24 hours column. In Per minute / 15 min, one number is the per-minute cap. Two numbers are the per-minute cap, then the per-15-minute cap. * These quotas are tuned per platform configuration and can differ from the value shown. Get Identity Limits always returns the number that actually applies.
Every window is its own cap. A shorter window stops a burst while the 24-hour quota still has room. A blank cell means that window has no separate cap.

Which limits each action consumes

Several actions draw on more than one limit, and hitting any one of them stops the action. 33 documented actions consume no per-action quota at all — they remain subject to the identity-wide caps and to LinkedIn itself.
Profile enrichment is served by several internal paths, each metered separately. The ceiling that applies to a given identity at a given moment is the one Get Identity Limits returns — not the table above.

Identity-wide throughput

Separate from the per-action quotas above, each identity also has a ceiling across all actions combined: See How Limits Work for how this relates to the other systems. Ramp-up does not change these caps.

Checking an identity’s real numbers

The tables above show full capacity. The Get Identity Limits endpoint returns what actually applies to one identity right now — ramp-up and any custom configuration included.
Evaluating limits in real time is expensive, and the call is rate limited alongside other identity operations. Query it when a limit surprises you — don’t poll it.

When every identity is exhausted

With identity_mode=auto, Edges spreads work across the identities available for the action. If all of them have reached their limit for that action, the run is postponed rather than failed:
Other actions on the same identities may still be running normally — limits are per action. Your options are to add identities, switch to identity_mode=managed and let Edges supply the capacity, or wait for postponed_until.

Avoiding LinkedIn restrictions

Smart Limits protect the account from volume. Two things outside our control matter just as much.

Don’t run other automation alongside Edges

Many tools handle IP addresses poorly. Running several at once means LinkedIn sees the same account acting from different places, which is one of the strongest signals it has for flagging automation. If a client’s account is restricted, this is the first thing to check.

Keep the timezone honest

The timezone on an Edges identity provisions a proxy IP in that country. If it doesn’t match where the person actually logs into LinkedIn — including their VPN, when they use one — LinkedIn sees an account logging in from two countries. Either match the timezone to the VPN, or have them disconnect the VPN for LinkedIn.
Restrictions are usually temporary. Work through the two checks above before assuming volume was the cause.

When a limit is reached

Every limit rejection carries a label saying which window was exhausted, so you can wait the right amount of time rather than backing off blindly.

Which limit was hit

Every label, its origin, and the right retry for each.

Recover a blocked run

Picking work back up after a run stopped.

How Limits Work

The five systems, and which ones you have to plan around.

Get Identity Limits

The effective numbers for one identity.

Live Mode Safety

Pacing live calls, which the quotas alone won’t do for you.