LinkedIn doesn't publish official limits. They never have. But after analyzing hundreds of account restriction reports across Expandi, LinkedHelper, Waalaxy, and HeyReach communities, clear patterns emerge. Here are the numbers that keep your account safe.
The Hard Limits (Never Exceed These)
| Action | Free LinkedIn | Premium | Sales Navigator |
|---|---|---|---|
| Connection requests/day | 20–25 | 25–30 | 30–40 |
| Messages/day (1st degree) | 50–80 | 80–100 | 100–150 |
| InMails/month | — | 5–15 | 20–50 |
| Profile views/day | 80–100 | 100–150 | 150–200 |
| Group messages/day | 20–30 | 30–40 | 40–50 |
| Likes/day | 50–80 | 80–100 | 100–150 |
| Comments/day | 20–30 | 30–40 | 40–50 |
These are daily limits. Not weekly averages. Not "some days more, some days less." Consistency matters more than volume.
The Real Risk Formula
LinkedIn doesn't just look at volume — they look at behavior patterns. Here's what triggers restrictions:
- Burst activity: 0 actions for 3 days, then 40 connection requests in 2 hours. This is the #1 trigger. LinkedIn's system flags accounts that go from dormant to hyperactive.
- Identical messages: Sending the exact same message to 50 people. Even with template variables, if 80%+ of your message is identical across recipients, the pattern detection triggers.
- Low acceptance rate: If you send 100 connection requests and fewer than 15 are accepted, LinkedIn concludes you're spamming strangers.
- High report rate: If 2+ people mark your message as spam within a 7-day window.
- New account + high volume: Accounts under 6 months old that exceed 50% of the limits above.
Safe Settings by Tool
Expandi (Cloud-Based)
Expandi is the safest tool because they enforce conservative limits. Recommended settings:
- Connection requests: 15–20/day
- Messages: 20–30/day
- Working hours: 9am–6pm in your timezone
- Random delays: Enabled (default)
- Days off: Saturday + Sunday
LinkedHelper (Desktop-Based)
LinkedHelper runs on your machine, so you control everything — which means you can exceed limits if you're not careful. Recommended:
- Connection requests: 15–20/day
- Messages: 25–35/day
- Enable "smart delays" (random 30–180 seconds between actions)
- Enable "working hours" (never run campaigns at 3am)
- Cancel pending requests older than 30 days (built-in feature — use it)
Waalaxy (Chrome Extension)
Waalaxy runs in your browser, similar safety profile to LinkedHelper.
- Connection requests: 20–25/day
- Messages: 30–40/day
- Use Waalaxy's "Safe Mode" settings
- Don't use the "Visit profiles before connecting" feature more than 40/day
HeyReach (Multi-Channel)
HeyReach adds email to the mix, which distributes risk across channels.
- LinkedIn connection requests: 15–20/day
- LinkedIn messages: 20–30/day
- Email sends: 50–100/day (separate deliverability rules apply)
- Stagger LinkedIn and email — don't send both to the same person on the same day
The Warm-Up Protocol (New or Restored Accounts)
If your account was recently restricted, or you're warming up a new account:
| Day | Connection Requests | Messages | Profile Views |
|---|---|---|---|
| 1–3 | 5/day | 5/day | 10/day |
| 4–7 | 10/day | 10/day | 20/day |
| 8–14 | 15/day | 15/day | 30/day |
| 15–21 | 20/day | 20/day | 40/day |
| 22+ | Full limits | Full limits | Full limits |
During warm-up: accept all pending invitations manually, respond to every message, post at least 2–3 times. Show LinkedIn that a real human is using the account.
What To Do If You Get Restricted
- Stop all automation immediately. Don't log in from a different IP. Don't try to "fix" it by being more active.
- Wait 24–48 hours. Most first restrictions are 24-hour cooldowns.
- Log in from your normal device/IP. Complete the identity verification if asked.
- Manually use LinkedIn for 7 days — post, comment, respond to messages. No outreach.
- Resume automation at 50% of your previous volume for 2 weeks.
- Third restriction = permanent ban risk. If you've been restricted 3 times, you need to fundamentally change your approach or switch to a different strategy.
How OpenHive Stays Inside the Limits
Most automation tools enforce caps as fixed numbers — set 20/day, the tool sends 20/day, every day, until LinkedIn flags you. The caps are right; the enforcement is dumb.
OpenHive runs adaptive daily limits as a built-in agent behavior:
- Per-account warm-up phase — new accounts start at 30% of the cap and ramp over 21 days, matching the warm-up schedule in this guide.
- Friction detection — agents watch for LinkedIn's friction signals (extra captchas, slowed page loads, "are you a robot" prompts, soft-paused inboxes) and back off automatically that day, even if you haven't hit your nominal cap.
- Human-pattern timing — sends spread across the day with weighted distribution against your actual workday, not a constant trickle that screams "automation."
- Restriction tracking — every account on your team surfaces in a single dashboard with its current health (green / yellow / red). The Reviewer agent stops auto-approving new campaigns on any account in yellow or red.
If you do get restricted, OpenHive's Auto-Accept agent pauses immediately — no further sends — and surfaces the incident with the last 24h of activity so you can diagnose the trigger.
The principle: safety is enforced by the agent layer, not by you remembering to throttle. If you're spending mental energy worrying about whether you're sending too much, you've already lost — the tool should worry for you.