Usage limits

Limits exist so that a runaway loop at one integrator does not degrade the service for everyone else. They are sized so that a normal company never meets them.

What is included

Limit	Value
Requests per calendar month	25,000
Requests per minute (burst)	60

The per-minute limit catches runaways — a loop with no stop condition, a badly written retry. The monthly limit is the one that matches your real usage.

Both limits apply per company and per mode. Test mode is counted separately: it has its own quota, of the same size, and never eats into your live quota. Trying things out costs nothing.

What happens as you approach the limit

We do not cut the service off at 100 %. A hard stop would turn a commercial conversation into a production outage, and it would hit precisely those companies whose business is growing.

Instead:

You are notified at 50 %, 80 % and 100 % of your monthly quota.
Past 100 %, an overage of 20 % is allowed — once per calendar month. Your calls keep going through.
Once that overage is spent, calls are refused until the next month.

The allowance does not renew within the same month: it is there to give you time to call us, not to let you settle above your limit for good.

When a limit is reached

The response is a 429, with the headers that say what to do:

HTTP/1.1 429 Too Many Requests
Retry-After: 30
Tay-RateLimit-Limit: 60
Tay-RateLimit-Remaining: 0
Tay-RateLimit-Reset: 1758186794

Two codes, because the two failures call for different reactions. rate_limited means slow down and retry; quota_exceeded means the month is spent and retrying will not help until it resets.

{
  "error": {
    "code": "rate_limited",
    "message": "Too many requests: the limit is 60 per minute. Retry after 30s, and see https://tayo.pro/contact if you need a higher limit."
  }
}

{
  "error": {
    "code": "quota_exceeded",
    "message": "The monthly quota of 25000 requests, and the overage allowance on top of it, are spent. It resets on 2026-10-01. See https://tayo.pro/contact to agree a higher limit."
  }
}

Both carry a Link: <…>; rel="help" header pointing at the page for asking for a higher limit. There is no self-serve upgrade — see below.

Respect Retry-After rather than retrying immediately. A client that retries a 429 in a loop only spends its own allowance.

Checking your usage

curl https://app.tayo.pro/api/v1/usage \
  -H "Authorization: Bearer tay_live_your_key"

{
  "mode": "live",
  "period": "2026-09",
  "resetAt": "2026-10-01T00:00:00.000Z",
  "limits": { "monthlyRequests": 25000, "burstPerMinute": 60, "source": "default" },
  "month": { "used": 12483, "remaining": 12517, "state": "within" }
}

state is within below the quota, overage inside the one-time allowance, and exhausted once calls are being refused. source reads override when your company has a negotiated limit rather than the published value.

You get the mode of the key presented: a tay_test_… key reports test usage, never the live figures.

This endpoint does not count against the monthly quota and is never refused by it: it is the one call that still answers once the quota is spent. It remains subject to the per-minute limit.

Counters are kept per calendar month, in UTC.

Usage, along with your failed requests of the last 30 days, is also visible in Settings → API.

If you need more

Contact us, from Help → Contact the team or at tayo.pro/contact. The limit is raised per company, with no change to your code, and we agree together on a price that matches your volume.

There is no tier grid to pick from: heavy usages differ too much from one another for a grid to do them justice.

This conversation is commercial, not support: it is never charged — see API support.