Skip to main content

gRPC

Overview​

Starting with Weaviate v1.19.0, a gRPC interface has been progressively added to Weaviate. gRPC is a high-performance, open-source universal RPC framework that is contract-based and can be used in any environment. It is based on HTTP/2 and Protocol Buffers, and is therefore very fast and efficient.

As of Weaviate v1.23.7, the gRPC interface is considered stable. The latest (v4) Weaviate Python client library supports gRPC, and the other client libraries will follow.

Protocol Buffer (Protobuf) definitions​

A gRPC interface is defined through its Protocol Buffer, or Protobuf (read more) definitions.

In the case of Weaviate, the .proto files are listed in the Core library's proto directory.

This directory contains the following files:

  • weaviate.proto: The main Protobuf definition file. This file defines the Weaviate service, and specifies the RPC methods available in the Weaviate service.
  • batch.proto: Defines data structures for handling batch object operations. This file is imported by weaviate.proto.
  • search_get.proto: Defines data structures for handling search (get) operations. This file is imported by weaviate.proto.
  • base.proto: Defines base data structures to be used elsewhere. This file is imported by batch.proto and search_get.proto.

How to use gRPC​

Server-side​

To take advantage of the gRPC API, we recommend using the latest version of Weaviate. We suggest using port 50051 for gRPC calls. This can be set in the configuration file for Weaviate. (Note that Weaviate Cloud Services (WCS) uses port 443 for gRPC.)

As an example, the snippet below maps 50051 as the host port so that it can be accessed from outside the container. The 50051 port is mapped to the 50051 port inside the container for gRPC calls, and the 8080 port is mapped to the 8080 port inside the container for REST calls.


```yaml
---
version: '3.4'
services:
  weaviate:
    # ... Other settings
    ports:
     - "8080:8080"  # REST calls
     - "50051:50051"  # gRPC calls
  # ... Other settings

Client-side​

You can use the gRPC interface through the (v4 Weaviate Python client library). gRPC support in the other client libraries will follow.

Alternatively, you can use other tools, such as the grpcurl command-line tool, to interact with the gRPC API. Some options include: