Lykke HFT API allows users to use two kinds of protocols: gRPC API
and Rest API
. Both APIs support the same methods and data format, however, only gRPC API
is able to receive streaming data from the platform.
The gRPC API
utilises a new generation of RPC framework that includes working with HTTP 2
, ProtoBuf
, and WebSocket
for faster and more efficient interaction with the platform. The gRPC framework uses the declarative description of the contract in the .proto files, on the basis of which it is simple to generate Client Library using numerous programming languages. Furthermore, the gRPC framework supports working with stream making it easy to receive events directly from the server.
Lykke team strongly recommends using gRPC API protocol to communicate with the platform
Read more about gRPC framework
: https://grpc.io/
A useful tool for manual gRPC API requests:
- BloomRPC - UI tool to interact with gRPC servers (when using BloomRPC make sure to use TLS with server certificate (TLS button))
- grpcurl - a command-line tool that lets you communicate with gRPC servers
API Endpoint:
hft-apiv2-grpc.lykke.com:443
- host format
Proto files
- common.proto :Common data models
- isalive.proto : Is Alive API service
- publicService.proto : Public API service
- privateService.proto : Private API service
Rest API
uses a classic HTTP based framework that entails working with HTTP 1.1
and JSON
. Rest API
allows users to call RPC methods without streaming data from the server.
Disclaimer: The rest API has been added for backward compatibility.
Lykke team recommends against using the Rest API protocol in favour of the gRPC protocol
.
API Endpoint:
A useful tool for manual rest API requests: HFT API Swagger
PUT
: Updates a resource.POST
: Creates a resource.GET
: Gets a resource or a list of resources.DELETE
: Deletes a resource.
- 200
OK
: The request was successful. - 401
Unauthorized
: Authentication failed. - 404
Not Found
: Endpoint was not found. - 500
Internal Server Error
: Server error.
Every response contains two fields - payload
and error
. A successful response will contain the response data in the payload
field and the null in the error
field and vice versa for the error response.
Here you have a list of errors you might encounter in the paragraph Error codes at the end of the document.
Successful response. Property
error
is null.
{
"payload": {
...
},
"error": null
}
package hft;
message Response {
PayloadType payload = 1;
hft.common.Error error = 2; // error is NULL
}
Error response. Property
error
is not null.
{
"error": {
"code": "itemNotFound",
"message": "Asset not found",
"fields": {
"assetId": "Asset not found"
}
},
"payload": null
}
package hft.common;
message Error {
ErrorCode code = 1; // 1100
string message = 2; // "Asset not found"
map<string, string> fields = 3; // "assetId" : "Asset not found"
}
enum ErrorCode {
success = 0;
runtimeError = 1001;
itemNotFound = 1100;
/* Full list in the paragraph **Error codes** at the end of the document */
}
package hft;
message Response {
PayloadType payload = 1;
hft.common.Error error = 2; // error is NOT NULL
}
You can create your API keys in here. You can also check our step by step guide here.
To use the API keys, you should just add a header Authorization: Bearer <your API Key>
with the bearer token on your request.
Request Header
"Authorization": "Bearer **********************************"
"Authorization": "Bearer **********************************"
Here you can see how to manage decimal
types (Price, Volume, Amount, etc) in API contract.
In the gRPC API
contract, the decimal
type is presented as a string
type, with a textual representation of the number. This is done in order to avoid issues with the non-strict precision double
type.
In the Rest API
contact, the decimal type is presented as a number with strict precision.
Example with decimal type
{
"price": 222231.33420001911,
"volume": 0.0000001
}
message Body {
string price = 1; // "222231.33420001911"
string volume = 2; // "0.0000001"
}
Here you can see: How to manage the TimeStamp
type in the API contract.
The timestamp is always used in the time zone UTC+0
In the Rest API
contact, the TimeStamp
type is presented as a number
with "Milliseconds in Unix Epoch" format of date-time.
In the gRPC API
contract, the TimeStamp
type is presented as a google.protobuf.TimeStamp
type.
Example with timestamp
{
"Timestamp": 1592903724406
}
import "google/protobuf/timestamp.proto";
google.protobuf.Timestamp time_name = 1; // 1592903724406
List of possible order states
Name | Meaning |
---|---|
Placed | Order in the Order Book. |
PartiallyMatched | Order in the Order Book and partially filled. |
Matched | Order is filled. |
Pending | Order is pending a trigger to be placed in the Order Book. |
Cancelled | Order is cancelled by the User. |
Replaced | Order is replaced (canceled) by the user. |
Rejected | Order rejected by the System. |
API has a request rate limits.
If client hit limit then API return error: 429 Too Many Requests
Rate Limit settings:
API type | methods | Request rate limits |
---|---|---|
gRPC API |
hft.PrivateService.* | 1000 request per minute per method |
gRPC API |
hft.PublicService.* | 300 request per minute per method |
Rest API |
/api/orders/* | 300 request per minute per method |
Rest API |
/api/* | 120 request per minute per method |