Show HN: Open SaaS – An open-source alternative to paid boilerplate starters

I too built my own boilerplate, Rails + blogging + Kamal deployments, and call it Business Class. Not free, but worth it:)

I can't spend more than $5 a month, and I don't want to use anything other than Python. I think in Python and it provides the absolute frictionless path to bring my ideas to life. I wish I had the same feeling about JavaScript, and that would have certainly made my life easier. But this is a compromise I am happy to accept as part of my identity.

The stack I am thinking of right now includes:

- Firebase: I chose this because Stripe integration, real-time database, Firebase functions, documentation and Firebase authentication.

- Vue/Nuxt + BootstrapVue: This is for basic front-end stuff.

- VM/VPS: This will be used to host the operational logic.

- Python: This is the core operation. Firestore operations and management will have the real database in PostgreSQL/SQLite to reduce calls to Firestore.

- FastAPI: This will be used for communicating with Firebase via API. User calls will be relayed via Firebase to FastAPI. It may be slower, but it will make my life a bit easier.

- NGINX: Plan is to have one FastAPI project per project. I think NGINX is going to be helpful here. Also, I would like to do some firewall configuration.

I still have a lot to figure out. One thing I am trying to determine is if I really need Firebase here because I already have a backend there.

I want to build an API-first system and Unix philosophy-inspired systems where multiple API services or Python modules do one thing only and are linked to each other. The goal is to create common operation Python modules, so I can quickly develop POCs by combining boilerplates with parameters and write the core logic in the minimum amount of Python code. The whole plan is to keep trying different approaches until I find one that works.

Mine is built on CUE, which at least has the potential to become a more widely used language. CUE hasn't reached sufficient maturity for broader adoption yet, so I continue to face this same problem.

https://github.com/hofstadter-io/hof

What we built is more like a framework for building WASP + boilerplate for anything, both sides of the transform in full control by the user.

I'm less bullish on the idea since LLMs that can code arrived on stage

Recently looked at templ-great features, but somewhat awful looking generated code.

It's using Golang + Chi + Templ, uses SQLite as its database and has a multi-tenant-multi-db setup by default (i.e. 1 master database for user/tenant management, and "stem-data", 1 sqlite database per tenant), and uses passkeys for authentication. I'm using some HTMX + Hyperscript + a really small amount of plain-js for passkeys in the apps themselves.

I reckon that switching out Chi for Fiber - for the boilerplate only - would take about 90 minutes.

Edit: Forgot to include TailwindCSS. So for completeness sake.

We'd definitely like to make it more flexible (also in the sense of UI frameworks, etc), but that's something we'll have to figure out step by step.

The problem is how it's labelled as a true generic solution instead of being a JS SaaS boilerplate. Developers often either have specialisations or preferences, and luckily we have more than one option for each kind of application.

See SaasPegasus for a good example. Clearly Django oriented. I have my own boilerplate for my work in Laravel, but I don't brand it as the new saas kid on the block.

No gripes against OP, looks like a neat setup. Just the messaging reads to me off. Like it's focusing on the OSS side of it in a weird way. Plenty of boilerplates exist out there that aren't paid solutions, and all stack their own preferred dev tools that are majority open source. It's as if this were a product called "peanut flavoured peanut butter".

- CoverLetterGPT (https://coverlettergpt.xyz/) - generate a cover letter based on your CV and job description

- Etsy description generator (this one even got acquired): https://dev.to/wasp/from-idea-to-exit-building-and-selling-a...

Close?

It's more a mental barrier as new tech no one has heard of is always somewhat intimidating, whether for good reason or not.

Also re OSS positioning - I'm actually not aware of modern (but maybe I missed it), polished SaaS starters such as the paid ones that got really popular lately (e.g. supastarter, shipfast), so that's what prompted us into that kind of messaging. We'd like to create a community-driven starting point for developers of equal polish and quality.

And that's exactly what we are going for, just with an open-source, community-driven approach.

I can imagine you had the same motivation when you created your boilerplate starter (not the oss/community thing, but above)?

On concurrency, I hope FastAPI will help on that. But I am horrible at writing concurrency code and that is why I don't enjoy writing JS. But from my limited experience, I enjoyed goroutines.

Package management, for mission critical stuff I have to straight up use Go binaries to be honest.

I didn't mention Go in my parent comment because it will be used as a langauge to rewrite solutions and for that to happen I need to make atleast a couple hundred dollars a month to make that investment.

But Wasp promises so much that I still want to try it. Getting auth and RPCs auto-generated would be great on its own.

Still, it looks cool despite not knowing React myself. I might take it for a spin when I have some free time :) thanks for building this

On a side note, LinkedIn has a lot of frequent posts if you follow certain projects. LlamaIndex is killing it on this front. Lots of great stuff and they even teamed up with Andrew Ng with advanced RAG techniques on deeplearning.ai

Calling it as easy as JSON for JS dev sounds a bit stretched. For a react Dev maybe.

Also first time I'm hearing of the term Djangonaut, love it! :)

People learn what they are eager to