Skip to main content

Overview

PromptGuard implements two types of limits to ensure fair usage and system stability:
  1. Monthly Request Quotas - Based on your subscription plan
  2. Rate Limiting - Maximum requests per minute (anti-abuse)

Monthly Request Quotas

Your subscription plan determines how much usage you get per month, as usage credit at list prices: These apply to every counted request, on every endpoint: the proxy, POST /api/v1/guard and POST /api/v1/security/scan alike. The full list of what counts is in Plans & Limits. Scans from the Shadow AI desktop agent count against the same quota but are never refused for it.

Hard vs Soft Limits

Free plan uses a hard limit:
  • When you have spent your $5 of usage for the month, requests return 429 Too Many Requests
  • Upgrade to Developer ($20 of usage a month) to continue
Developer plan uses a soft quota with a hard spend cap:
  • When you have spent your $20 of usage credit, requests continue processing
  • Past the credit, Requests are billed at the same list prices, on by default
  • You receive email alerts at 80%, 90% and 100% of the credit (how quota warnings work)
  • Spending is bounded by your cap (default **$100**, up to $10,000); past it requests return 429 spending_limit_exceeded until you raise it
  • Set the cap to $0 to stop at the credit instead of paying overage
Example (Developer plan):

Checking Your Usage

View current usage in the dashboard:
Or via API:

Rate Limiting

PromptGuard enforces per-plan, per-account rate limits on all /api/v1/* endpoints: These limits apply to your whole account; adding more API keys does not increase them, and neither does creating more organizations. The account is the one that pays: every organization you own shares your plan’s limit, and a teammate’s keys in your organization get your plan’s limit, whatever plan they are on themselves. The one exception is a client organization a managed service provider created (a Managed Organization, which has no owner of its own): each one gets its own budget, at its provider’s plan limit, so one client cannot spend the minute of the provider’s other clients or of the provider’s own keys. Once the client has an owner of its own, its keys spend that owner’s budget like any other organization’s. The budget refills at the start of each minute. Scans from an enrolled Shadow AI device neither count toward it nor are refused by it. Separately, a Cloud Armor anti-abuse layer enforces a per-IP request limit at the network edge. The per-IP limit is independent of your plan and exists to block abusive traffic.

Per-key limits

Because the plan’s limit is shared by every key in the account, one key that leaks or runs away can spend the whole minute and leave your other keys refused. To contain that, give a key a requests-per-minute limit of its own: Dashboard → API Keys → ⋯ → Rate limit. In a team organization only its admins can set or remove it, including on a key a member created; in your personal organization you can. Every change is recorded in the audit log as api_key_rate_limit_changed, with the limit before and after.
  • The key is held to both limits, and whichever is lower refuses first. A per-key limit never raises your plan’s.
  • A request refused by the key’s own limit does not count toward the account’s budget, so your other keys keep theirs.
  • It takes a whole number of at least 1; a value above the maximum is refused with a message that names it. Leave the field empty to remove the limit, and the key is held to the plan’s alone.
  • It counts in the same one-minute window as the plan’s limit and refills at the same moment. It is enforced on self-hosted installs as well, where the plan limits are not.
  • Rotating the key carries its limit to the new key.
  • A key created over the API can be given one with rate_limit_per_minute (Create API key), and must be when the calling key has one.
  • An enrolled Shadow AI device’s key cannot take one, because device scans are never rate limited (409 device_credential_rate_limit). A teammate who is not an admin is refused with 403 api_key_rate_limit_admin_only.

Rate Limit Headers

Every /api/v1/* response carries the limit the key is held to, what is left of it this minute, and when the window resets:
All three describe the same limit, so Remaining counts down from Limit and refills at Reset. A 429 with Retry-After is still the signal to stop.

Handling Rate Limits

If you exceed your plan’s per-minute rate limit, or a key’s own, you’ll receive a 429 Too Many Requests response:
The type and code are the same whichever limit refused; the message says which (your plan allows … or this API key allows …). retry_after (and the Retry-After header) is the number of seconds until the minute ends and your budget refills. Recommended handling:

Idempotency Keys

For safe retries of POST/PUT/PATCH requests, include an Idempotency-Key header:
If you retry the same request with the same idempotency key within 24 hours, you’ll get back the cached response with an X-Idempotency-Replayed: true header. This prevents duplicate operations.
Idempotency keys are scoped to your API key and expire after 24 hours.

Best Practices

1. Implement Exponential Backoff

2. Monitor Usage Proactively

Set up monitoring to alert before you hit limits:

3. Batch Requests When Possible

Instead of:
Use batch processing:

4. Cache Responses

Cache frequently requested results:

Upgrading for Higher Limits

Need higher rate limits or custom quotas? Enterprise plans offer:
  • The highest per-minute limit (1,000/min)
  • Custom monthly request quotas
  • IP allowlisting for API access control
  • Idempotency keys for safe retries
  • Support and availability terms agreed in your contract
Contact us at sales@promptguard.co for Enterprise pricing.

Frequently Asked Questions

Why do different plans have different rate limits?

Rate limits scale with your plan tier (Free: 60/min, Developer: 600/min, Enterprise: 1,000/min). These are per-account limits, except that each client organization a managed service provider created has its own budget at the provider’s limit (see Rate Limiting). The Cloud Armor per-IP limit is a separate anti-abuse layer at the network edge.

What happens if I consistently go over my monthly quota?

On Free, requests are blocked with 429 errors until you upgrade. Developer keeps processing past its credit at list prices, up to your spend cap; past the cap requests return 429 until you raise it. Enterprise continues processing per contract. On every plan you are emailed at 80%, 90% and 100% of the quota, subject to your Email Alerts setting, and on a paid plan the dashboard shows a banner from 90%; Reaching your limit has the details.

Can I increase my rate limit?

By moving to a plan with a higher limit. Each plan’s per-minute limit is fixed, and there is no per-organization override. You can set a lower limit on a single key (see above) to stop one key spending the whole budget.

Do retries count against my quota?

Yes. Every request to our API counts, including retries. Implement smart retry logic with exponential backoff to minimize wasted quota.

How is usage calculated?

One request = one call to a counted endpoint (the proxy, /api/v1/guard, /api/v1/security/scan or /api/v1/security/redact — the full list is in Plans & Limits), regardless of:
  • Number of tokens
  • Response length
  • Model used

Monitoring Tools

Dashboard Analytics

Track usage in real-time:
  • Current period usage
  • Daily/weekly/monthly trends
  • Over-quota events
  • Rate limit hits

Usage API

Programmatically monitor usage:
Returns:

Need Help?


Next steps

Error handling

Every error code the API returns and how to handle it

Usage tracking

Allowance used, spend this cycle, tokens per model

Pricing

Plan limits, quotas and how usage is counted
Last modified on November 13, 2026