Limits & caps
CloudyBot meters usage with hard monthly caps (and some daily caps). When a meter is exhausted, the related feature pauses — no surprise overage bills on most plans. Full tier table: pricing. Billing and resets: billing & plans.
Hard-cap behavior
- Pause, don’t sneak-charge — chat may return a limit error; browser sessions stop starting; sends are blocked.
- Per-meter — exhausting browser minutes does not always mean AI credits are gone; each feature has its own counter.
- Reset restores full allowance — unused capacity does not roll over (see billing FAQ for Free vs paid reset timing).
- Overflow wallet — after monthly AI credits are used, optional bonus task credits (referrals, credit packs) can extend chat; browser and sends still obey their own caps.
Main meters (by feature)
Source of truth: api/features/billing/plans.js. Highlights for Free vs paid patterns:
- AI credits (
taskCap) — charged per chat/tool turn complexity. Free: 100/month. Growth: 3,000. Pro: 6,000. Max: 18,000. - Browser minutes — cloud browser automation time. Free: 30/month. Scales up on paid tiers (750 → 3,500).
- Web searches — dedicated search tool quota. Free: 30/month.
- WhatsApp proactive sends — outbound messages the agent initiates. Free: 0. Paid tiers include 100–500/month.
- Specialists (
maxEmployees) — Free: 2 (Cloudia + one custom; preset templates need Growth+). Growth: 5. Pro: 10. Max: 50. - Scheduled duties (
cronJobs) — Free: 2. Growth: 6. Pro: 15. Max: 50. - Chat history retention — Free: 30 days stored; Max: unlimited. Older messages may purge per plan.
- Model picker — Free and Growth: Auto-only. Pro and Max: manual model selection including premium models on Max.
Secondary meters
- AI image generation (
imagesPerMonth) — separate from branded social templates. Free: 3/month. - Social post images (
socialImagePostsPerMonth) — Sharp template renders. Free: 0; Growth: 300; Pro/Max: effectively unlimited (enforced as 9,999). - Code execution — sandbox runs per month (Free: 5).
- SEO audit runs — included on Pro (3/period) and Max (20/period); separate from the one-time Audit Pack product.
- Email sends — daily cap per account (dashboard shows sends remaining today); independent of AI credits.
What typically counts
- AI credits — assistant replies, tool orchestration, and multi-turn continuations in chat or cron.
- Browser minutes — active cloud browser session time, not idle dashboard viewing.
- Failed runs — partial work may still consume credits or minutes depending on how far the run progressed; split large jobs across scheduled chunks instead of one megaprompt.
- Cron chains — the first turn of a scheduled run still executes; continuation turns may soft-stop with a 429 if credits are already exhausted mid-chain.
Tips to stay within caps
- Scope prompts narrowly — see sending messages.
- Use scheduled tasks to spread browser work instead of one marathon session.
- Deploy multi-Specialist workflows only when your plan has enough Specialist and cron headroom.
- Upgrade or buy credit packs from Settings when you consistently hit AI credit limits before month end.
Common questions
What happens when I hit a limit?
The related feature pauses until reset, upgrade, or overflow credits (AI tasks only). You should see a clear error in chat or the dashboard — not a silent charge.
Can Free users send WhatsApp?
Not proactively. Free includes 0 outbound WhatsApp sends per month. Inbound/setup may work after connection, but agent-initiated sends need a paid plan with quota left.
Do limits apply per thread or per account?
Per account. All threads share the same monthly meters. Thread memory is isolated; billing is not.
Where do I see usage?
Open the dashboard — usage summaries appear in Settings and contextually near capped features (for example email sends remaining today).
Related: FAQ hub · blog: what hard caps mean · contact