Why safe rates are non-negotiable
Every social platform has anti-spam thresholds. They're undocumented (platforms won't publish them, that'd help spammers). They change without notice. Cross them and the consequences range from invisible (shadowban for 7-30 days) to permanent (account ban with no appeal).
Mirgent's defaults sit well below every documented spam threshold we know of. The numbers below come from production observation across thousands of Mirgent users; treat them as floors not ceilings.
Scheduling
Posts/day per platform
Lead Discovery
Replies/day per platform
AI Agents
Follows/day per platform
Analytics
Random delays between actions
Posting rate limits per platform
Conservative ceilings; Mirgent enforces these by default.
Twitter / X
Posts: 25/day safe, 50/day yellow, 100+/day red. Threads count as one post. Replies: 50/day safe, 100/day yellow. Follows: 50-100/day safe (varies by account age).
Posts: 5-10/day safe (LinkedIn rewards quality over quantity). Connect requests: 15/week safe, 30/week yellow, 100/week red (free accounts). Likes: 100/day safe.
Posts: 3-5/day safe (less if account < 3 months old). Comments: 5-10/day safe per subreddit, 50/day total. Self-promotion: 1 promotional post per 10 organic comments (the 1:10 rule).
Account age matters. New Twitter / Reddit / LinkedIn accounts (< 30 days) should run at 50% of these limits. Established accounts (1+ year) can safely raise to 150%. Mirgent's defaults are calibrated for established accounts.
Engagement rate limits
Likes, follows, comments, all rate-limited.
Likes
Twitter: 100/day safe. LinkedIn: 100/day safe. Instagram: 200/day safe. Reddit: unlimited (no spam threshold for upvotes/downvotes).
Follows + unfollows
Twitter: 25-50/day each direction safe. Instagram: 50/day each. LinkedIn: 15/week (free) or 30/week (premium). Don't exceed 5x normal-user volume regardless of account age.
Comments + replies
Reddit: hardest cap (5/day per subreddit). Twitter: 30-50/day. LinkedIn: 30/day. Instagram: 50/day. All assume context-relevant comments; spam patterns flagged regardless of count.
Mirgent's enforced defaults
Out-of-the-box, Mirgent stays well below every cap. Defaults you can review and tune in Settings → Rate Limits.
Random delays
Mirgent inserts 30-90 second random delays between actions. Looks human; prevents burst patterns that trigger anti-spam.
Per-action daily caps
Per platform, per action type. Posts: 25/day Twitter, 5/day LinkedIn. Replies: 5/day Reddit, 30/day Twitter. Hard caps; no override on Free / Creator plans.
Cooldowns
Per-thread cooldown (don't reply to same thread twice within 24h). Per-user cooldown (don't engage with the same user >5 times/week). Per-subreddit cooldown (Reddit only, 24h).
When to raise limits (and when not to)
Most users never need to raise default limits. Specific cases where raising is justified:
Scheduling Guide
Calendar, kanban, and queue modes
AI Agents Guide
Configure autonomous AI workflows
Integrations
Connect RSS feeds and platforms
Lead Discovery Guide
Find leads on Twitter and Reddit
Pitfalls in rate management
Mistakes that turn safe automation into bans.
- Raising all limits at once. If you must raise, do it one at a time and observe for a week. Mass raise = your account stops resembling normal users overnight, triggers anti-spam.
- Ignoring per-thread cooldowns. Replying to the same thread 3x in a day on Reddit gets you instant ban regardless of comment quality.
- Running multiple Mirgent workspaces against the same platform connection. Each workspace operates within its own caps; in aggregate you exceed thresholds. Use separate platform connections per workspace.
- Posting at burst times. Even within daily limits, 10 posts in 1 hour reads spammy; spread across the day. Mirgent's optimal-time scheduling enforces this automatically.
If you hit a rate limit
Recovery patterns by severity.
Yellow flag (warning, not blocked)
Stop the relevant action category for 48-72 hours. Resume at half-rate. Most yellow flags clear within a week of reduced activity.
Shadowban (your posts don't reach feeds)
Pause all automation for 7-14 days; engage manually only. Most shadowbans lift in 14-30 days. Resume Mirgent at 30% of previous rates after lift.
Hard ban / suspension
Appeal through platform's normal channel. Mirgent doesn't help with appeals (we have no special access). If appeal fails, account is gone, start fresh on a new account, more conservative defaults this time.
{guide_safe_walk4_title}
{guide_safe_walk4_desc}
{guide_safe_walk5_title}
{guide_safe_walk5_desc}
Safe Rates FAQ
How does Mirgent know what the safe rates are?
Production data from thousands of Mirgent users. We track which rate configurations produce zero issues vs which produce occasional yellows. Defaults reflect the never-issued-warning band; raising above pushes into yellow probability.
Can I disable rate limits if I'm willing to risk it?
On Pro+ plans, yes, per-action limits are configurable. On Free / Creator, no (defaults are mandatory). The plan distinction is intentional: serious operators understand the risk; new users shouldn't be able to foot-gun their accounts in the first week.
What if I'm running 5+ accounts on one platform?
Each connected platform-account in Mirgent has its own rate budget. 5 Twitter accounts = 5 separate 25-posts-per-day budgets. Just don't run them from the same IP, platforms cluster-detect IP patterns and treat them as one.
Are these rates the same on the API as via the extension?
API rate limits are typically much lower than the human-equivalent caps Mirgent uses. That's why the extension model gets higher safe ceilings, you're operating at human user limits, not API tier limits.
Why does Mirgent ship with so-conservative defaults?
Banning an account is far more expensive (lost reputation, lost followers, lost history) than running slightly under maximum throughput. We optimize for not-getting-banned over getting-the-most-throughput-this-week. The math is: raising 20% throughput risks 100% of your account.
{guide_safe_faq6_q}
{guide_safe_faq6_a}