APIReference

Limits

Rate limits, one conversion at a time, and file limits.

Markdown llms.txt

4 min read

Coming soon

The Developer API opens soon. You can read the docs now; API keys arrive when it opens.

On this page

Every account has the same limits. If you need more, email contact@perfectvector.com: we can raise them for your account.

Request limits

Limit	Live keys	Test keys	When you reach it
Conversions created per minute	20	20	`429 rate_limited`
`POST /v1/vectorizations` attempts per minute (every `POST`, also replays and refused requests)	about 40	about 40	`429 rate_limited`
Other requests per minute (status, downloads, `DELETE`, `GET /v1/account`)	about 120	about 120	`429 rate_limited`
Live conversions in progress at once	1	no limit	`429 too_many_in_flight`
Active API keys	10, live and test together		`409` in the Developers area

Limits apply to the whole account, across all of its keys. Test keys have their own counters.
Creates are counted over the last 60 seconds: the next one is allowed when the oldest is a minute old. With test keys you can send 20 at once; live keys also run one conversion at a time (below). A replay with the same Idempotency-Key does not count as a create.
POST attempts count every POST, including replays and requests refused with an error. The limit is twice the create limit, so it only matters for clients that retry in a tight loop.
Attempts and other requests are counted by each of our servers separately, so these limits are approximate. Other requests allow a burst of up to 30 at once, then the rate of about 120 a minute.
One conversion at a time. A second live POST while one is running answers 429 too_many_in_flight with Retry-After: 5, or 402 insufficient_credits when your available credits cover only the running conversion. Send the next image when the first one has finished. Requests are not queued for you.
Busy service. When the API as a whole is at capacity, a POST answers 503 service_busy with a Retry-After. Retry with the same Idempotency-Key.
Failed authentication. Many requests with invalid keys from one IP address get 429 rate_limited for a few minutes.

Rate-limit headers

Authenticated responses carry the limit the request counted against: for a POST that created or replayed a conversion, the create limit; for a POST refused before that, the attempts limit; for everything else, the limit for other requests.

Header	Meaning
`X-RateLimit-Limit`	Requests a minute for this limit.
`X-RateLimit-Remaining`	Requests you can send right now. For other requests this is at most the burst of 30.
`X-RateLimit-Reset`	Seconds until the next request is allowed again under this limit (for creates: until the oldest create in the window is a minute old).
`Retry-After`	On `429` and `503`: seconds to wait before the next try.

Every 429 and 503 problem body also carries retry_after (seconds), except too_many_in_flight, which uses the header only.

Responses from the edge

Our network edge can answer some requests before they reach the API, with a 403 or a 429 whose body is not application/problem+json and has no request_id. Slow down and retry later. If it keeps happening, contact us with the time of the request and your IP address.

File limits

Limit	Value
Formats	PNG, JPEG, WebP (checked from the file's content, not its name)
File size	5 MiB (5,242,880 bytes)
Image size	100 to 4096 pixels on each side
Frames	One. Animated images are refused.
Request	`multipart/form-data` with the field `image` (required) and the field `options` (optional). No other fields.

GET /v1/account returns your current limits in limits:

Example request

"limits": {
  "requests_per_minute": { "create": 20, "read": 120 },
  "max_in_flight": 1,
  "max_variants": 1,
  "max_file_bytes": 5242880,
  "min_dimension": 100,
  "max_dimension": 4096,
  "formats": ["png", "jpeg", "webp"]
}

With a test key, max_in_flight is null.

Reviewed October 30, 2026