## Which plan you need

The public API (and the own MCP server) are available **only on the Pro and Enterprise plans**. On the
Starter plan, any call to `/public/v1/*` responds `403 PLAN_NOT_ELIGIBLE` — even if the key exists, isn't
expired or revoked, and has the correct scope. It's a server-side check on **every** request, not only when
the key is created: if your account drops a plan tier, your existing keys stop working immediately (nothing
needs to be deleted), and they work again as soon as you upgrade.

## Request limits

Every key has its own quota, tiered by the plan of the tenant it belongs to:

| Plan | Per minute | Per day |
| --- | --- | --- |
| Starter | — (no access, see above) | — |
| Pro | 180 | 50,000 |
| Enterprise | 600 | 200,000 |

Going over either ceiling returns `429 RATE_LIMIT_EXCEEDED`. These numbers are a starting point (not yet
measured against real production traffic) — if your integration needs more sustained quota, write to
**hola@nuntius.chat**.

## Rate-limit headers

Every `/public/v1/*` response includes your remaining quota:

```
X-RateLimit-Limit: 180
X-RateLimit-Remaining: 179
X-RateLimit-Reset: 1735689600
```

`X-RateLimit-Reset` is a Unix timestamp (seconds) — when your quota resets. A `429` also adds `Retry-After`
(seconds until you can retry):

```bash
curl -i https://api.nuntius.chat/public/v1/leads -H "X-API-Key: aa_YOUR_FULL_KEY"
# HTTP/1.1 429 Too Many Requests
# X-RateLimit-Limit: 180
# X-RateLimit-Remaining: 0
# X-RateLimit-Reset: 1735689600
# Retry-After: 42
```

The quota is counted **per key** (by the key's hash, not by your IP address): rotating IPs doesn't free up
quota, and two different keys on the same tenant have independent quotas.