How Badgr protects your LinkedIn account

Account safety isn’t a marketing claim or a setting you can toggle. It’s a set of operational decisions baked into how Badgr runs every action on your behalf. There are four pillars; this article walks through what each one actually does.

Pillar 1: Dedicated proxies, one per account

Every LinkedIn account you connect to Badgr gets its own residential proxy IP. Two important things:

Not shared between customers. Your proxy isn’t pooled with other Badgr customers’ accounts. LinkedIn never sees two different accounts using the same IP.
Not rotated. The proxy is “warm” — it stays assigned to your account for the life of the connection. LinkedIn sees a stable IP, the same way it would for a real person.

We use residential proxies (real consumer ISP IPs), not datacentre proxies. Datacentre IPs are cheap, but they’re trivially identifiable by LinkedIn’s anti-abuse system.

The proxy is geo-matched to where the account holder is actually based. A London-based founder gets a London IP. Mismatches between account location and IP location are themselves a flag.

Pillar 2: Per-account dynamic throttling

LinkedIn’s “weekly invitation limit” isn’t a fixed 100 like the internet folklore says. It’s dynamic, per-account, and lives somewhere between 80 and 200 depending on:

How old the account is
The acceptance rate of recent invitations
Whether the account has Sales Navigator
Recent reply rates and connection activity
A few other signals LinkedIn doesn’t publish

Badgr reads each account’s individual ceiling from observed signals and stays under it. A new account might be at 60/week while an established one is at 180. The system adjusts continuously — if LinkedIn starts throttling your account, our cap drops automatically before the account hits a hard restriction.

See Per-account throttling explained for the deeper version.

Pillar 3: Behaviour-randomised actions

Real humans don’t send 30 invitations at 9:00:00. They don’t operate 24 hours a day. They don’t send identical message text to 50 prospects in the same minute.

Badgr’s behaviour engine adds:

Randomised micro-delays between actions (10 seconds to 3 minutes)
Session gaps — not every minute of every day, but clusters of activity with realistic breaks
Working-hours windows scoped to the account holder’s timezone, not server time
Variable per-step pacing — not 50 connection requests in the morning then 0 for the day

The point isn’t to fool a human observer. It’s to make the activity pattern indistinguishable from “engaged human user” rather than “scheduled bot.”

Pillar 4: Hard kill-switch on verification challenges

LinkedIn occasionally pushes a security verification — a captcha, a phone code, an “is this you?” prompt. These aren’t punishments by themselves; they’re sanity checks.

The wrong thing to do is retry. Many automation tools see a verification, fail the action, log “error”, wait 30 minutes, then try the same action again. That’s the fastest way to escalate from “we want to verify you” to “your account is restricted.”

Badgr halts the specific account the moment a verification prompt appears. No retries, ever. The account stays paused. You complete the verification in LinkedIn directly (we’ll notify you). Once cleared, Badgr resumes from where it stopped. Other accounts in your workspace are unaffected.

What we deliberately don’t do

A few practices common in the category that we won’t ship:

Shared proxy pools. Cheaper, but the single biggest cause of multi-account restrictions.
Datacentre IPs. Trivially fingerprinted as automation.
Retry-on-failure for verification prompts. Compounds the signal that the account is bot-driven.
Sales-side claims about restriction rate that we can’t yet prove. We’ll publish our actual restriction-rate vs the industry once we have a full year of operational data.

If you want the full technical brief — proxy provider, exact throttle parameters, action-log schema — we share it under NDA with prospective customers. Email hello@badgr.ai.