ONLINE 11/2025 Personal

grpc-gen

Scaffold a full-stack gRPC + GraphQL gateway + business-rule engine from .proto files in 3 commands

GogRPCProtobufgqlgenGraphQLMongoDBMySQLRedisDocker

GitHub pkg.go.dev

Commands → full-stack

Architecture layers

Integration tests

MIT

Open source

Overview

A code generator written in Go, inspired by gqlgen, Hasura and PostgREST: you write .proto files and the tool emits an entire production stack — gRPC services, a GraphQL gateway in front of them, and a MongoDB-backed business engine for validation, row-level security and multi-step compute actions; all wired with TLS, structured logging, DataLoaders and Dockerfiles. The engine is data-driven (rules/scopes/actions are Mongo documents, hot-reloaded without a rebuild). The emitted code is readable Go you commit to your repo, with no runtime dependency on the tool. Production-tested against two real apps (an LMS and a thesis-management system) with ~700 integration tests each.

Role: Author & maintainer — designed the CLI, code generator and business engine

System architecture

> Client

Web / Mobile

GraphQL over HTTP

> Gateway

gqlgen

schema + resolvers

DataLoaders

Redis cache

@auth

MongoEnforcer

Admin REST

runtime CRUD

> Business Engine

business_rule

validation

read_scope

RLS

action

pipeline

> Services (gRPC / mTLS)

CRUD Handlers

MySQL

Filter whitelist

Tracer + logger

> Storage

MySQL

MongoDB

engine state

Redis

MinIO

file uploads

Modules & Services

CLI

initadd-serviceadd-gatewayproto gen

Services

gRPC CRUDMySQL poolFilter whitelistPaginationDockerfile

Gateway (gqlgen)

GraphQL schemaResolversDataLoaders@auth directiveAdmin REST

Business Engine

business_ruleread_scope (RLS)action pipelineHot-reload

Infra

MySQLMongoDBRedisMinIOmTLS

Developer UX

Readable GoNo runtime depStructured loggingCross-platform binary

Key features

3 commands to scaffold the whole system: init → add-service → add-gateway

Generates gRPC CRUD services: MySQL pool, filter whitelist, pagination, Dockerfile

GraphQL gateway via gqlgen: schema + resolvers + DataLoaders (Redis 2-layer cache) + @auth directive

Data-driven business engine on MongoDB: business_rule (validation), read_scope (RLS), action (post-mutation pipeline) — hot-reload

Runtime Admin REST API to CRUD rules / policies / scopes / actions

mTLS between gateway ↔ services, function tracer + structured logging

Emits readable Go committed to your repo, with no runtime dependency on the tool

Cross-platform binaries: linux / macOS / windows × amd64 / arm64

Highlights

Inspired by gqlgen, Hasura & PostgREST — generates readable Go, not a runtime framework

Production-tested on two real apps, ~700 integration tests each

Data-driven engine: edit rules/RLS/actions as Mongo docs + reload, no rebuild

Three cleanly separated layers: Services / Gateway / Engine, deployed independently