Why this page has numbers on it
Because the alternative is a promise. "Human-like behaviour", "smart limits", "LinkedIn-safe" — every competitor writes some version of that, and none of it can be checked. A ceiling you can read is a claim you can hold us to.
Two things are true at once, and this page is written to hold both:
- The mechanism is real, specific, and mostly not built by anyone else at this depth.
- Nobody, us included, can publish an honest restriction rate. LinkedIn does not release one, no vendor has a control group, and any number in either direction is invented. So what follows is a description of what the software does — not a promise about what LinkedIn will do.
The daily ceilings
These are the fully warmed limits — what an account is allowed once it has finished the ramp. Per account, per day.
| Action | Booost ceiling | Where the number comes from |
|---|---|---|
| Connection requests | 10 / day, 50 / week | LinkedIn publishes 100/week. This is exactly half. |
| Messages | 50 / day | No published cap; vendors claim 50–150. We take the low end. |
| Profile visits | 40 / day | No published cap; 80/day widely repeated. Half of it. |
| Reactions | 25 / day | A minority source gives 50/day on Free and Premium. Half of the stricter reading. |
| Comments | 15 / day | Claims run 50–100 for established accounts. We sit far under. |
| Posts | 5 / day | No cap exists, but a 2M-post study finds a ~40% reach penalty on a same-day second post. |
| Searches | 10 / day | Governed by a monthly budget, not a daily one. |
The rule behind the table: about half of the best-evidenced practical threshold, and where sources disagree the safer reading wins — not the average. Where there is no evidence at all, we sit at or below the most conservative serious peer.
Two of those numbers came down in July 2026 rather than up. Profile visits went from 100 to 40, because it was the one action where we sat above prudent peers on a real ban vector — sweeping profiles is what list-building bots do. Posts went from 50 to 5, because a ceiling only reachable by behaving like a bot protects nobody.
The warm-up: 42 days, five phases
A new or newly connected account does not start at those ceilings. It starts at 15% of them and climbs, on calendar days from the moment the account is connected.
| Phase | Days | Share of the ceiling |
|---|---|---|
| 0 · Soft start | 0–7 | 15% |
| 1 · Light | 8–14 | 30% |
| 2 · Ramp | 15–28 | 55% |
| 3 · Consolidate | 29–42 | 80% |
| 4 · Full | from day 43 | 100% |
Phase 0 is a real trickle, not a pause: the account does a little, visibly, from day one. Full speed arrives on day 43.
The warm-up state is per account, stored locally, and it does not reset because you restarted the app. ⚠️ It does reset if you reinstall on a clean machine without a restore — which is the whole point of #219 and the reason a backup exists.
The trust score
A number per account, recomputed from measured signals. It modulates the limits around the warm-up: when it dips, the system slows itself down without being asked.
| Factor | Weight | Measured from |
|---|---|---|
| Acceptance rate, 30 days | 25% | tracked connection requests |
| Incidents, 14 days | 20% | refusals, CAPTCHAs, rate limits |
| Profile completeness | 15% | the account's own LinkedIn profile |
| Warm-up phase | 15% | where the account is on the ramp |
| Interaction quality (NDI) | 15% | observed interactions |
| SSI score | 10% | LinkedIn's own number, when extractable |
The part worth knowing: factors with no real data are excluded and the remaining weights are renormalised. The score is never padded with a neutral placeholder to look complete. If we have not measured your acceptance rate yet, acceptance rate does not silently score you 50.
That design came out of a bug worth admitting. The score used to include account age, which measured how long the row had existed in Booost's own database — so a LinkedIn veteran with ten years of history scored zero on it the day they installed the app, surfaced as "Needs care", and could not exceed 70 for months. Profile completeness answers the question account age was pretending to ask — is there a real, established person behind this account — and answers it on day one.
What stops an account
Four independent brakes. Any one of them is enough.
1The circuit breaker
A CAPTCHA, a rate limit, an unexpected response shape — automation stops for that account and resumes on its own once things look normal again. It is not a setting you can forget to switch on.
2Human pacing, in the account's own timezone
Randomised delays between actions, confined to the operating hours you set. It does not work at 4am, and it does not fire twenty actions in a minute and then sleep.
3Auto-pause on reply
The moment a prospect answers, the sequence stops and hands you the conversation. This is a safety feature and a decency feature at once: nothing is more obviously automated than a follow-up that ignores the reply above it.
4The human veto
Every action a campaign wants to take can be refused by a person before it happens. The inbox drafts replies rather than sending them.
The part that is structural, not configurable
All of the above is settings and arithmetic. This is not:
The actions come from your own computer, in a browser window that belongs to the application, using a session that never leaves your disk. A cloud tool logs into LinkedIn from a data-centre address shared with every other customer on that platform, on infrastructure LinkedIn can characterise in one query.
We argue that mechanism — whose IP, whose session, whose machine — and we do not attach a number to it, because there isn't an honest one. See how the data actually flows for what does and does not leave your machine.
Safety is not the paid part
The full anti-ban system is in all three editions, including the free Community one. There is no "safety pack", no premium limits tier, and no cheaper plan that runs your accounts harder.
The only ceilings that differ between editions are Community's publishing quotas — one post a day, five comments, ten reactions — and those are commercial, not protective. Everything on this page is identical whether you pay us nothing, €79 or €99. See the editions in full.
Questions this page gets asked
Can I raise the limits? Not above the ceilings on this page. The warm-up and the trust score move you within them.
Does the warm-up restart if I stop using an account? No. The phase is calendar-based from connection, and it does not roll back for inactivity.
What happens if LinkedIn restricts an account anyway? The circuit breaker trips and automation stops. Booost cannot lift a restriction — nobody's tool can — and any page suggesting otherwise is worth closing.
Do you have a ban rate to show me? No, and neither does anyone else honestly. See the top of this page.
Does this work for Company Pages too? Yes, with their own separate limits. ⚠️ One known gap: two admins of the same Page running the same workflow each act under their own account, and the Page has no shared memory of it — tracked as #278.