Rate limits

Rate limits protect the platform from abuse and keep latency predictable. They apply in layers: IP / edge limits for anonymous tools, workspace scan cadence for product scans, and plan quotas for how much of the product you can use.

Free tools and quick scan

Public free tools (schema, robots.txt, metadata, FAQ helpers, crawlability checks, and similar) and quick-scan endpoints are rate limited per client IP. Limits use a sliding window; exceeding them returns a retry-later response.

• Tool endpoints share a tools bucket (on the order of tens of requests per minute per IP).
• LLM-backed helpers use a stricter bucket.
• Attribution collect traffic uses a higher-throughput bucket appropriate for pageview beacons.
• When distributed Redis is unavailable, limits fall back to per-instance memory (weaker under multi-instance load).

Product scan cadence

• Authenticated workspaces also have a soft anti-abuse cap on scan creation per hour (shared across manual and monitoring starts).
• This is separate from monthly or lifetime plan entitlements.
• If you see a “scan limit reached / hour” message, wait and retry — it is not an upgrade wall by itself.

Plan quotas and entitlements

Product quotas are plan-based. Free and paid plans differ on completed scans, how many recommendations/weaknesses are visible, history depth, competitors, monitoring cadence, projects, and included AI credits.

• Free workspaces typically include a limited lifetime completed-scan allowance.
• Paid plans raise or remove scan caps and unlock fuller report surfaces, subject to metering.
• Entitlements sync from Polar subscriptions; past-due states can fall back to free limits.
• See Billing and Manage billing for plan details.

FAQ