
A full-stack starter with Go/Gin, React 19, auth and multi-tenancy already done — so you can jump straight into business logic.
Every time I start a new project, I do the same thing.
Set up the Go backend. Set up the React frontend. Figure out how they talk to each other in dev. Figure out how they deploy together. Add auth. Add sign up, sign in, sign out. Protect the routes. Store the token somewhere. And then , finally , I can start working on the thing I actually wanted to build.
That part is always the same. And it's always slow.
So I searched on GitHub for a starter that could save me this work. I found a lot of repos, but most of them had one of these problems:
I just wanted one repo where I could clone it, set two env vars, run make dev, and start writing my feature. That repo didn't exist the way I wanted it. So I built it.
Here it is: github.com/calebeaires/gin-react-monorepo
It's a full-stack monorepo with:
You clone it, set DATABASE_URL and AUTHULA_SECRET, run make dev, and you already have sign up, sign in, protected routes, organizations, members, roles, and tenant-scoped endpoints working.
You sign up, you get a token, the token goes in a Bearer header, the middleware checks it on every request. No magic. No third-party dashboard. The auth code is in the repo and you can change it.
This is the part I hate doing from scratch. Every B2B app needs organizations. Users belong to orgs. Each org has its own data. Roles. Invites. All of it.
The repo uses schema-per-tenant on Postgres. Each org gets its own schema (tenant_<slug>). When a request comes in with the X-Org-Slug header, the middleware starts a transaction and sets search_path to that tenant's schema. Your handler doesn't know or care , it just queries as normal and the data is isolated.
Creating an org creates the schema. Deleting an org drops it. Owners, admins and members have different permissions. It's all already there.
In development, Vite runs on :5173 and Go runs on :8081. Vite proxies /api and /auth to the Go server, so everything is same-origin and you don't fight CORS.
In production, Vite builds the React app to static files, Go embeds them with //go:embed, and you ship one binary. One file, one process, one port. No Nginx. No separate static server. No Docker Compose with three services.
The README has a full tutorial for adding a "Projects" feature end to end , model, migration, handler, route, frontend page. Six steps, tenant-isolated, authenticated, done.
That's the whole point. The boring stuff is done. You just add the thing you care about.
If you're like me and you want to build a B2B SaaS or any app with accounts and organizations, and you don't want to spend the first week of every project wiring up the same boilerplate, this is for you.
If you like Go but don't want to pick a frontend stack from scratch every time, this is for you.
If you want multi-tenancy done the right way without reading ten blog posts about search_path, this is for you.
This is open source (MIT) and I want it to grow. I built the first version for myself, but a starter like this only gets really good when more people use it and push back on the decisions.
So if this sounds useful to you, here's how you can help:
make dev, build something small with it. Tell me what broke, what felt weird, what was missing.I want this to be the repo I wish I had found on GitHub. I can't get it there alone.
Link: github.com/calebeaires/gin-react-monorepo
Thanks for reading. Now go build the thing you actually wanted to build.
aiMost of us have seen a coding agent fail to complete a task we know it can do. We just don't...
googlecloudWhen building Generative AI applications, developers often encounter a massive bottleneck: sequential...
discussI’ve been thinking about sharing some electronic circuit posts on Dev.to — small circuits, DIY...
agentsWhat nobody tells you about exporting your multi-agent prototype to a local workspace. Every...
agenticarchitectAutonomous agents are genuinely good at answering messy business questions. Give one an LLM and a set...
aiPR volume went up, ticket quality didn't, and the gap got filled with LLMs on both sides of the review: bots reviewing, bots replying, bots occasionally arguing with bots about priorities that only existed in a teammate's head. Our CEO named the actual problem, and it's bigger than code review.
Workflows from the Neura Market marketplace related to this Stable Diffusion resource