Framework Comparison

How Go Micro compares to other Go microservices frameworks.

Quick Comparison

vs go-kit

go-kit Philosophy

• "Just a toolkit" - minimal opinions
• Compose your own framework
• Maximum flexibility
• Requires more decisions upfront

Go Micro Philosophy

• "Batteries included" - opinionated defaults
• Swap components as needed
• Progressive complexity
• Get started fast, customize later

When to Choose go-kit

• You want complete control over architecture
• You have strong opinions about structure
• You're building a custom framework
• You prefer explicit over implicit

When to Choose Go Micro

• You want to start coding immediately
• You prefer conventions over decisions
• You want built-in service discovery
• You need pub/sub messaging

Code Comparison

go-kit (requires more setup):

// Define service interface
type MyService interface {
    DoThing(ctx context.Context, input string) (string, error)
}

// Implement service
type myService struct{}

func (s *myService) DoThing(ctx context.Context, input string) (string, error) {
    return "result", nil
}

// Create endpoints
func makeDo ThingEndpoint(svc MyService) endpoint.Endpoint {
    return func(ctx context.Context, request interface{}) (interface{}, error) {
        req := request.(doThingRequest)
        result, err := svc.DoThing(ctx, req.Input)
        if err != nil {
            return doThingResponse{Err: err}, nil
        }
        return doThingResponse{Result: result}, nil
    }
}

// Create transport (HTTP, gRPC, etc)
// ... more boilerplate ...

Go Micro (simpler):

type MyService struct{}

type Request struct {
    Input string `json:"input"`
}

type Response struct {
    Result string `json:"result"`
}

func (s *MyService) DoThing(ctx context.Context, req *Request, rsp *Response) error {
    rsp.Result = "result"
    return nil
}

func main() {
    svc := micro.NewService("myservice")
    svc.Init()
    svc.Handle(new(MyService))
    svc.Run()
}

vs gRPC

gRPC Focus

• High-performance RPC
• Multi-language support via protobuf
• HTTP/2 transport
• Streaming built-in

Go Micro Scope

• Full microservices framework
• Service discovery
• Multiple transports (including gRPC)
• Pub/sub messaging
• Pluggable components

When to Choose gRPC

• You need multi-language services
• Performance is critical
• You want industry-standard protocol
• You're okay managing service discovery separately

When to Choose Go Micro

• You need more than just RPC (pub/sub, discovery, etc)
• You want flexibility in transport
• You're building Go-only services
• You want integrated tooling

Integration

You can use gRPC with Go Micro for native gRPC compatibility:

import (
    grpcServer "go-micro.dev/v6/server/grpc"
    grpcClient "go-micro.dev/v6/client/grpc"
)

svc := micro.NewService("myservice",
    micro.Server(grpcServer.NewServer()),
    micro.Client(grpcClient.NewClient()),
)

See Native gRPC Compatibility for a complete guide.

vs Dapr

Dapr is a distributed application runtime. Its building blocks cover service invocation, state, pub/sub, bindings, secrets, configuration, distributed locks, actors, jobs, and workflow, usually accessed through a sidecar from many languages. Dapr Agents adds an agent framework on top of those runtime capabilities.

Go Micro overlaps with Dapr on distributed-systems primitives, but the product shape is different: Go Micro is a Go framework where services, agents, tools, and flows are built from the same runtime. A service endpoint can become an AI-callable tool, and an agent is itself a registered service with memory, guardrails, planning, delegation, MCP, and A2A around it.

Decision table

When to choose Dapr

• You need a polyglot runtime contract for Node, Python, Java, .NET, Go, and other services.
• Your platform team already operates sidecars and component configuration across Kubernetes clusters.
• You want Dapr's standard building blocks for state, pub/sub, bindings, secrets, actors, jobs, and workflow more than you want a Go-native service framework.
• You are adopting Dapr Agents and want to stay in its Python-first agent stack.

When to choose Go Micro

• You are building mostly in Go and want the agent harness to be the same runtime as your services.
• You want service methods and their comments/examples to become AI-callable tools without maintaining a separate tool layer.
• You want agents to be deployed, discovered, called, load-balanced, and inspected like ordinary services.
• You need MCP and A2A at the agent/service boundary, not only an internal application API.
• You prefer library-native composition and direct Go interfaces over sidecar component wiring.

Where Go Micro still needs to prove itself

Dapr has a mature platform narrative and broad deployment footprint. Go Micro's agent-harness story is sharper for Go teams, but production adoption depends on keeping the no-secret getting-started path green, documenting durability semantics clearly, proving MCP/A2A conformance with external clients, and making Kubernetes deployment first-class.

Practical migration path

1. Start with one Go Micro service that wraps a real domain capability.
2. Add doc comments and examples so the endpoint is useful as an agent tool.
3. Expose it through MCP for external agents or through A2A if the capability is itself an agent.
4. If your platform already uses Dapr, keep Dapr components behind the service boundary and let Go Micro present the agent/tool contract.
5. Move deterministic multi-step work into flows only after the service/tool boundary is stable.

vs Agent Frameworks (Google ADK)

ADK (Agent Development Kit) is Google's open-source, code-first framework for building AI agents. It spans several languages (Python, TypeScript, Go, Java, Kotlin); adk-go is the Go implementation. It's model-agnostic (optimized for Gemini), speaks MCP and A2A, and supports multi-agent systems, evaluation, and deployment to Cloud Run / GKE.

They overlap on agents but solve different problems. ADK is a library for building an agent process — you define an agent, its tools, and a model, then run and deploy it. Go Micro is the harness around agents once they operate real systems: service discovery, inter-service RPC, pub/sub, durable flows, tool execution, and deployment. Those pieces are out of scope for ADK, and you bring your own.

In Go Micro an agent is built as an ordinary service: it registers in the registry, is callable by RPC (Agent.Chat) and over A2A, and other services and agents discover and call it the same way they call anything else. Its endpoints are exposed as MCP tools automatically. So once you have more than one agent or service, Go Micro also gives you the discovery, RPC, pub/sub, config, and deployment around them.

When to choose ADK

• You want an agent framework with first-class evaluation and context tooling
• You're polyglot, or invested in the Google Cloud / Gemini ecosystem
• You want a cross-language A2A ecosystem with Google's backing

When to choose Go Micro

• You want an agent harness where agents and services are the same thing — registered, discoverable, load-balanced, and deployed the same way
• You want your existing services to become agent tools with zero extra code (every endpoint is an MCP tool automatically)
• You're building in Go and want one set of primitives for services, agents, and flows

They interoperate

Both speak MCP and A2A, so this isn't strictly either/or: a Go Micro agent and an ADK agent (in any language) can call each other over A2A, and either can consume the other's MCP tools. A common pattern is to run Go Micro as the service mesh / runtime and let ADK (or any A2A agent) plug into it.

vs tRPC-Agent-Go

tRPC-Agent-Go (maintained by tRPC-Group, validated inside Tencent) is a production-grade Go framework for agent systems: LLM / Chain / Parallel / Cycle / Graph agents, function tools, MCP, A2A, AG-UI, Redis memory and RAG, evaluation, agent self-evolution, and OpenTelemetry. It's a serious, well-resourced project.

They overlap heavily on agents but take a different approach. tRPC-Agent-Go is an agent SDK you run alongside your services — you compose agents and tools into graphs and conditional workflows, and your microservices (tRPC) live separately and are called into. Go Micro starts from the premise that an agent is a service — one runtime where every endpoint is automatically a tool, an agent registers and is discovered and load-balanced like anything else, and workflows are durable code paths rather than a graph DSL. The premise is that the line between "your services" and "your agents" is accidental complexity; remove it and there's less to wire and keep in sync.

When to choose tRPC-Agent-Go

• You want a graph/workflow DSL for composing agents and tools
• You're on tRPC, or want to add agents alongside an existing service stack
• You want first-class evaluation and self-evolution today, with a large team behind it

When to choose Go Micro

• You want one runtime where services, agents, and flows are the same primitives — registered, discoverable, and deployed the same way
• You want your existing services to become agent tools with zero extra code
• You prefer durable flows and plain code paths over a graph DSL, in a small, independent framework you can hold in your head

They interoperate

Both speak MCP and A2A, so a Go Micro agent and a tRPC-Agent-Go agent can call each other over A2A, and either can consume the other's MCP tools. You can run Go Micro as the service-and-agent runtime and still reach an agent built on tRPC-Agent-Go.

Feature Deep Dive

Service Discovery

Go Micro: Built-in with plugins

// Zero-config for dev
svc := micro.NewService("myservice")

// Consul for production
reg := consul.NewRegistry()
svc := micro.NewService("myservice", micro.Registry(reg))

go-kit: Bring your own

// You implement service discovery
// Can be 100+ lines of code

gRPC: No built-in discovery

// Use external solution like Consul
// or service mesh like Istio

Load Balancing

Go Micro: Client-side, pluggable strategies

// Built-in: random, round-robin
selector := selector.NewSelector(
    selector.SetStrategy(selector.RoundRobin),
)

go-kit: Manual implementation

// You implement load balancing
// Using loadbalancer package

gRPC: Via external load balancer

# Use external LB like Envoy, nginx

Pub/Sub

Go Micro: First-class

broker.Publish("topic", &broker.Message{Body: []byte("data")})
broker.Subscribe("topic", handler)

go-kit: Not provided

// Use external message broker directly
// NATS, Kafka, etc

gRPC: Streaming only

// Use bidirectional streams
// Not traditional pub/sub

Migration Paths

See specific migration guides:

• From gRPC

Coming Soon:

• From go-kit
• From Standard Library

Decision Matrix

Choose Go Micro if:

• ✅ Building Go microservices
• ✅ Want fast iteration
• ✅ Need service discovery
• ✅ Want pub/sub built-in
• ✅ Prefer conventions

Choose go-kit if:

• ✅ Want maximum control
• ✅ Have strong architectural opinions
• ✅ Building custom framework
• ✅ Prefer explicit composition

Choose gRPC if:

• ✅ Need multi-language support
• ✅ Performance is primary concern
• ✅ Just need RPC (not full framework)
• ✅ Have service discovery handled

Choose Dapr if:

• ✅ Polyglot services
• ✅ Heavy Kubernetes usage
• ✅ Want portable cloud abstractions
• ✅ Need state management

Performance

Rough benchmarks (requests/sec, single instance):

Benchmarks are approximate and vary by configuration

Community & Ecosystem

• Go Micro: Active, growing plugins
• gRPC: Huge, multi-language
• go-kit: Mature, stable
• Dapr: Growing, Microsoft-backed

Recommendation

Start with Go Micro if you're building Go microservices and want to move fast. You can always:

• Use gRPC transport: micro.Transport(grpc.NewTransport())
• Integrate with go-kit components
• Mix and match as needed

The pluggable architecture means you're not locked in.