gRPC

Dynamic gRPC mocking — no compiled .proto files needed. Uses a raw codec to intercept any service/method call. Supports unary RPCs as well as server-streaming, client-streaming, and bidirectional-streaming.

protocols:
  grpc:
    enabled: true
    port: 50051
    services:
      - mocks:
          # Unary
          - id: charge
            method: Charge
            response:
              success: true
              charge_id: ch_123

          # Server-streaming: reads the client's one request, then sends
          # each `stream` entry in order (each with its own optional delay).
          - id: list-charges
            method: ListCharges
            stream:
              - response: {charge_id: ch_1}
              - response: {charge_id: ch_2}
                delay: 100ms

          # Client-streaming: drains every message the client sends, then
          # replies once with `response` after the client half-closes.
          - id: sum-charges
            method: SumCharges
            client_stream: true
            response:
              total: 42

          # Bidirectional streaming: client_stream + stream together reply
          # to each client message with the correspondingly-indexed stream
          # entry; once `stream` is exhausted, further client messages are
          # drained silently until it half-closes.
          - id: echo-charges
            method: EchoCharges
            client_stream: true
            stream:
              - response: {echo: 1}
              - response: {echo: 2}

Any stream entry may set error instead of response to end the call early with a gRPC status, same shape as the top-level error field.

By default, a streaming RPC (server- or bidirectional-streaming) returns once its stream list is exhausted. Set keep_open: true on the mock to leave it open indefinitely instead — including across a bidirectional call's client half-close — so further ad-hoc messages can be pushed through the stream management API.

Optionally reference a real .proto file (proto: ./protos/payments.proto) to generate mocks from it instead of hand-writing them — see Generating mocks from an OpenAPI, AsyncAPI, or Protobuf spec.

See also

• Configuration — matching, templated responses, state conditions, and scenarios shared by every protocol
• Fault Injection — this protocol's fault fields: code, message, delay, error_rate
• Management API — manage open streaming connections (push/close)