Rate Limit

RESTful Rate Limit

The specific rules of REST rate limit are as follows:

1. Resource Pool:

Each API resource pool has a certain quota, the specific amount of which depends on the VIP level:

Level	Unified Account	Spot (include Margin)	Futures	Management	Earn	CopyTrading	Public
VIP0	200/3s	4000/30s	2000/30s	2000/30s	2000/30s	2000/30s	2000/30s
VIP1	200/3s	6000/30s	2000/30s	2000/30s	2000/30s	2000/30s	2000/30s
VIP2	400/3s	8000/30s	4000/30s	4000/30s	2000/30s	2000/30s	2000/30s
VIP3	500/3s	10000/30s	5000/30s	5000/30s	2000/30s	2000/30s	2000/30s
VIP4	600/3s	13000/30s	6000/30s	6000/30s	2000/30s	2000/30s	2000/30s
VIP5	700/3s	16000/30s	7000/30s	7000/30s	2000/30s	2000/30s	2000/30s
VIP6	800/3s	20000/30s	8000/30s	8000/30s	2000/30s	2000/30s	2000/30s
VIP7	1000/3s	23000/30s	10000/30s	10000/30s	2000/30s	2000/30s	2000/30s
VIP8	1200/3s	26000/30s	12000/30s	12000/30s	2000/30s	2000/30s	2000/30s
VIP9	1400/3s	30000/30s	14000/30s	14000/30s	2000/30s	2000/30s	2000/30s
VIP10	1600/3s	33000/30s	16000/30s	16000/30s	2000/30s	2000/30s	2000/30s
VIP11	1800/3s	36000/30s	18000/30s	18000/30s	2000/30s	2000/30s	2000/30s
VIP12	2000/3s	40000/30s	20000/30s	20000/30s	2000/30s	2000/30s	2000/30s

2. Weight:

When a user requests any API, the weight of this interface will be deducted and updated every 30s (starting from the arrival time of the user's first request). For specific interfaces, please refer to the rate limit weight regulations under each interface.

If the quota of any resource pool is used up within 30s, that is, after the rate limit is exceeded, an error message of http code:429, error code:429000 will be returned, and the request can be re-requested after the length of time displayed in the request header. At this point, the user must stop access and wait until the resource quota is reset before continuing to access.

For example:

When the user's VIP is 5, s/he has a "total spot quota" of 16000/30s.

The quota consumption for each "add spot limit order" is 2. After placing the first order, the user's remaining spot quota is 15998, after placing the second order, the remaining quota is 15996, and so on.

If the quota is not used up within 30 seconds, when the next cycle comes, the spot resource pool quota will be reset and returned to the quota limit of 16000.

3. Request Header:

The returned information of each request will carry the following information: Total resource pool quota, resource pool remaining quota, resource pool quota reset countdown (milliseconds).

"gw-ratelimit-limit": 500

"gw-ratelimit-remaining": 300

"gw-ratelimit-reset": 1489

4. Public Endpoint Rate Limit

It is based on IP rate limitation. If there is a lot of demand for the use of public interfaces, it is recommended to use the Websocket interface instead of the REST interface (if the interface supports it). To avoid IP rate limit issues, you may:

Use one server to bind multiple IP addresses (IPv4 or IPv6).

Use different IPs to avoid IP rate limit issues.

5. Private Endpoint Rate Limit

With the exception of Public resource pool endpoints, all other resource pools are based on UID, and the request header will carry the rate limit information of the resource pool, such as:

Remaining rate limit times

Rate limit cycle time

The rate limits of the sub-account and the master account are independent of each other at the API request level; that is to say, if the demand for such interface access is relatively large, it can also be solved by using the sub-account.

Server Overload Rate Limit

In addition to the regular rate limit, server overload may also trigger the rate limit. After the rate limit, the error code is 429000, but the request header will not carry other personal rate limit information. This type of rate limit does not count toward the number of rate limits. It is recommended to try again later.

Requesting Higher Limits

If you are a professional trader or market maker and need a higher limit, please send your KuCoin account, reason and approximate trading volume to api@kucoin.com:

KuCoin account information

Reason for the request

Estimated trading volume

WebSocket Rate Limits

1. Maximum Concurrent Connections

API	Limit	Scope	Note
Classic API	≤ 800 concurrent connections	Private (authenticated) endpoints are counted by UID; Public (unauthenticated) endpoints are counted by IP	Master and sub-accounts are completely independent (different UIDs)
Pro API	≤ 512 public + ≤ 512 private connections per IP (1024 total)	Counted by IP; public and private connections are counted separately

2. Connection Establishment Rate

API	Limit	Note
Classic API	No limit
Pro API	150 new connections per 5 minutes per IP	If exceeded, the server may reject new connection requests. Reuse existing connections; avoid frequent disconnect/reconnect.

3. Client-to-Server Messages (Client → Server)

API	Limit	Scope / Notes
Classic API	100 messages per 10 seconds	Counted per single connection
Pro API	100 messages per 10 seconds	Counted per single connection; includes `subscribe`, `unsubscribe`, `ping`. WebSocket cancel-order messages are excluded. If exceeded, the server may disconnect the connection.

API	Max Topics per Request	Scope
Classic API	100	Counted per single connection
Pro API	No limit	N/A

5. Maximum Subscribed Topics per Connection

API	Product Line	Limit	Note
Classic API	Spot / Margin	≤ 400 topics
Classic API	Futures	No limit
Pro API	All	≤ 200 topics	If more topics are required, split across multiple connections

6. Usage Recommendations (Pro API)

Reuse existing WebSocket connections whenever possible.

Control the client message sending rate (≤ 100 messages per 10 seconds per connection).

Split topics across multiple connections when subscribing to a large number of topics.

RESTful Rate Limit#

1. Resource Pool:#

2. Weight:#

3. Request Header:#

4. Public Endpoint Rate Limit#

5. Private Endpoint Rate Limit#

WebSocket Rate Limits#

1. Maximum Concurrent Connections#

2. Connection Establishment Rate#

3. Client-to-Server Messages (Client → Server)#

4. Subscribe / Unsubscribe Requests#

5. Maximum Subscribed Topics per Connection#

6. Usage Recommendations (Pro API)#