SDKs & protos
There is deliberately no vendored SDK: the streams are standard gRPC, so the gRPC tooling your team already uses is the SDK. Reflection is enabled on the endpoint, and we provide the .proto files, so you can generate stubs in any language gRPC supports.
Proto packages
Four packages on one endpoint (stream.hyperliquidrpc.com:443), two of them wire-compatible with providers you may already use:
Reflection
The endpoint describes its own services, methods, and message types. That is useful for exploration with grpcurl, and for clients (like @grpc/proto-loader or Python's reflection support) that can build stubs without a codegen step:
# the endpoint describes itself, no proto files neededgrpcurl -H 'x-api-key: YOUR_KEY' stream.hyperliquidrpc.com:443 list # inspect a service and its request/response messagesgrpcurl -H 'x-api-key: YOUR_KEY' stream.hyperliquidrpc.com:443 \ describe hyperliquid_swaps.v1.SwapStreamingGenerate stubs
For production we recommend generated stubs from the .proto files, which we provide with your console account: typed messages, no runtime schema fetch.
pip install grpcio grpcio-tools # generate message classes + stubs from the provided protospython -m grpc_tools.protoc -I protos \ --python_out=. --grpc_python_out=. \ protos/hyperliquid_swaps.proto # then:# import hyperliquid_swaps_pb2 as swaps_pb2# import hyperliquid_swaps_pb2_grpc as swaps_pb2_grpcThe same commands apply to any of the four proto files; swap hyperliquid_swaps.proto for the package you need. Working client code using these stubs is in the quickstart.
Already have QuickNode or Dwellir stubs?
Method-by-method coverage of every package is in the API reference.