The safest LinkedIn automation tool publishes its caps

PoliteReach publishes its real limits: 12-16 jittered invites/day, a rolling 7-day cap under LinkedIn's ~100/week, an optional 3-week warm-up ramp, auto-withdraw at 60 days.

Every tool in this category claims to be the safest LinkedIn automation tool on the market. In August 2026 we read the live sites of the seven biggest to find their actual daily invite caps. None publishes one. You get 'smart limits', an 'advanced safety algorithm', a hero promising 3X meetings, and no number anywhere. PoliteReach is for founders and small sales teams running outreach on their own accounts. The account at risk is yours. This page lists every cap, delay, and kill switch we run, with the numbers attached, so you can audit us before you connect anything. Then hold us to them.

None of the seven biggest LinkedIn automation vendors publishes its daily invite cap as a number; every one of ours is on this page.

Invites go out at 12 to 16 a day, jittered

The risky action on LinkedIn is the connection invite, not the message, so the invite is what we cap. Per account:

• 12 to 16 invites a day, drawn fresh every morning; the long-run average is about 14
• A rolling 7-day counter that keeps each account under roughly 100 invites a week
• Sending only inside office hours you configure
• A warm-up ramp for new or dormant accounts, starting well below the cap and building over about three weeks

LinkedIn doesn't publish its weekly invitation ceiling. We stay under the widely observed one rather than pretend to know the exact line. And nothing goes out at 3 a.m. on a Sunday.

A flat cap is a signature. An account that sends exactly 25 invites every single day, weekends included, at a steady pace, looks like software because it is software. Jitter breaks that. PoliteReach draws a fresh number between 12 and 16 each day: a 13-invite Tuesday, a 16 Wednesday, a 12 Thursday.

Messages to people already connected to you carry no volume cap in PoliteReach. That's deliberate. Messaging your own network is ordinary LinkedIn use. The invite is what gets accounts restricted, so the invite carries the limits.

Everything that sends was written before the campaign

PoliteReach generates no text at send time. Every connection note and every message in the Day 1, 3, 7, and 14 sequence is written and approved up front, by you or by your AI with your sign-off. Nothing improvises. What the preview shows is what the prospect gets. And the sequence has a hard stop: the moment someone replies, their follow-ups halt for good. Nobody receives a Day 7 nudge under the answer they sent you on Day 4.

The browser works at human speed

PoliteReach types each message key by key, with a delay on every keystroke. It pauses a randomized beat before opening a profile, before it starts typing, and before it hits send. The gaps between one action and the next are randomized too. This is slower than pasting a message into a field in one shot. Slow is the point. Uniform timing is an easy bot signal, and a browser that does everything in constant time is uniform by definition.

A security checkpoint stops that account immediately

When LinkedIn shows an account a checkpoint, PoliteReach's circuit breaker trips. Every queued action for that account stops, with no retries and no waiting an hour to poke the page again. Automation stays down until you step back in. The breaker is per account: one flagged profile doesn't pause the others.

Invites still pending at 60 days get withdrawn

A pile of ignored pending invites is the signature of a spray-and-pray account, and it drags down your sent-to-accepted ratio. PoliteReach withdraws any invite still pending after 60 days, automatically and by default. Tighten the window if you like. The floor is 7 days. You can't accumulate a thousand stale invites by forgetting about them.

A real browser runs in the cloud, nothing injected

PoliteReach is not a Chrome extension. It runs a full headful Chromium in the cloud and injects no code into LinkedIn's page; it clicks and types where you would. No extension, no page hooks. Linked Helper makes the same no-injection argument for its desktop app, and it's a fair argument. The difference is where the browser lives: in the cloud, each account gets one persistent device profile, so LinkedIn sees the same device every session instead of whichever laptop you opened that day. The fingerprint never changes.

Only one browser runs at a time. Actions across all your accounts take turns instead of firing in parallel.

IP discipline is specific too: one residential or ISP IP per account, never shared, never datacenter. Our docs tell you to run any candidate proxy through ipapi.is and reject it unless it reports is_datacenter false and an ASN type of isp. Datacenter means no. 'Dedicated' is a billing term; it says nothing about what kind of network the IP sits on.

LinkedIn has no API for acceptance. PoliteReach infers it by diffing your connections list against your still-pending invites. Every tool that reports acceptances is doing some version of this, whatever the dashboard implies.

The seven biggest vendors publish adjectives, not caps

The seven have real strengths. Expandi is mature, with a large user base, agency features, and image and GIF personalization we don't offer. La Growth Machine has the best safety section of the seven: 100% cloud, 5G mobile proxies, and a stated eight years of limit data. Real credit for that. Linked Helper is the cheapest of the majors, and its no-code-injected claim is real. Most of the seven offer a free trial; PoliteReach doesn't yet. Then comes the number that matters.

One published number does exist: Waalaxy pitches up to 800 invitations a month on its top plans. That's roughly 185 a week against an observed platform ceiling of about 100. A pitch that exceeds the platform's own ceiling is a volume pitch, not a safety pitch. And HeyReach's ban wasn't a user account; LinkedIn removed the company's own page. A vendor's safety story should at minimum cover itself.

You can audit any vendor in half an hour

Run every vendor on your shortlist through this list, including us.

1. Ask for the daily and weekly invite caps in writing. 'Smart limits' is not a number. If sales can't email you a figure, the real cap is whatever you type into the settings box.
2. Ask what happens on a LinkedIn security checkpoint. The only good answer: automation on that account stops immediately and stays stopped until a human clears it.
3. Get the proxy IP they'll assign you and check its ASN at ipapi.is. Reject anything showing is_datacenter true or an ASN type of hosting, whatever the plan calls it.
4. Search the vendor's name plus 'banned' and 'restricted'. That search is how you'd learn LinkedIn banned HeyReach's own company page in March 2026.
5. Ask whether stale pending invites are withdrawn automatically, and after how many days. A tool that lets pending invites pile up forever is ignoring your sent-to-accepted ratio.

A vendor that won't put its caps in writing doesn't have caps; it has defaults you can override.

PoliteReach answers 1, 2, 3, and 5 on this page. On 4 we have no history to search, for better and worse. That gap is why the caps are published. With no track record to show, the numbers are what we can put up front.

PoliteReach is in early access and incomplete

PoliteReach is built by Diverse Labs and does one thing on your own account, carefully: personalized invites at 12 to 16 a day, then a prefilled Day 1, 3, 7, and 14 sequence that stops the moment someone replies. That's the product.

The seven vendors above have more features and more history than we do. None of them publishes how many invites their tool sends in a day. We just did. If you'd rather pick an outreach tool on published numbers than on adjectives, the waitlist is open.