Show HN: BackendLab – Code-free web app builder | Hacker News Reader

Show HN: BackendLab – Code-free web app builder

65 pointschris14095758 comments

Hi all,

I'm currently validating an idea for a new BaaS solution which will allow you to build and host a web app backend without writing a single line of code.

The app is called BackendLab and the landing page is here: https://www.backendlab.app/

The idea is that it allows you to define your database models by specifying what fields to include and BackendLab automatically creates a CRUD API which you can start using straight away.

We've just launched the public beta and are looking for early adopters (free of charge, just looking for some feedback on the tool)

Looking forward to hearing your feedback!

Thanks,

Chris

View on HN

Why only backend? You are competing with other solutions out there that have a frontend available directly from db schema. Some examples: https://www.salesforce.com/products/platform/overview/ https://github.com/sferik/rails_admin

I am not totally clear who your target customers are. I think backend alone will be a barrier for a non-technical person. For a developer, as said by others, CRUD/authentication usually isn't the most time consuming part.

With that said, congrats on shipping! I know this is a work in progress. I think your landing page needs more work/content too. Good luck!

There is a patent troll who owns creating a frontend from a database. He's successfully sued Microsoft and Oracle and put a code generation product out if business. So I'd tread carefully there.

Interesting, thanks for the heads up. But is this likely to be an issue for a UK-based startup? I understand its a bigger problem in the US

Probably not if you're UK based.

Postgrest[1] is similar, and seems very popular, so this sort of thing is scratching somebody's itch.

[1] https://github.com/PostgREST/postgrest

This needs someone with db skills (yet works well for someone without BE dev skills) though.

Hay Chris,

Looking like a great project, simple and fun.

Keep in mind that other similar projects tried the same approach before, namely firebase and parse.com. both for only so much traction because at the end, real world applications need more features - business logic, backend that uses other APIs, etc.

Still, looks like a fun project

I tend to agree - great prototype, but most coders whip out basic CRUD code quickly, and their time is really spent on all the additional features of an app. Often, changing the basic CRUD code in the process.

This could work fine as a rapid prototyping tool, but from a feature parity perspective, it looks like you end up with something equivalent to a Salesforce app or Sharepoint list (also non-code CRUD solutions), without the stigma and cost of enterprise platforms, but also without the integration with everything else you may do.

Hi Chris,

I think a product like this deserves to have a great design. I worked on the color scheme, copy, and spacing for you and came up with this: https://imgur.com/a/UUoEij5

The I replaced the hero image with is: https://unsplash.com/photos/trYl7JYATH0

Good luck with your project! I'm not in the market for your product right now, but if I am some day, I'll definitely test this out!

Cheers, David

Thanks David, this is interesting. I like that hero image, but not sure if the overall design might be a bit too dark.

Kind of interested to hear how anybody else in this thread rates David's version over mine?

For me, parent's design definitely feels more premium than the original, which to me seems a bit amateurish.

Here are my observations:

* Layout is too wide. Parent has made it more narrow which feels more natural to me.

* Form inputs are too wide. Same thing as above.

* Weird color palette. I like parent's color simplicity. Notice how he has used different shades of the same color.

* Too random images. IMO the images used for the hero and pricing sections don't make any connection with the product. Parent has left out these images and it is much nicer.

Takeaway, prefer simplicity if you're not good at design.

I tried to also improve the line height, make the headings and subtitles simpler, and align some elements that weren't on a grid. I think you could benefit by having a designer friend look it over for you.

>The idea is that it allows you to define your database models by specifying what fields to include and BackendLab automatically creates a CRUD API

I don't see the advantage over Firebase FireRules , AWS AppSync , Azure Cosmos or MongoDB Stich all of which have advanced querying options with great SDK and integrations support.

This idea is not bad , it's just it doesn't seem to bring much compared to others on the market already offering BaaS...

Hi, thanks for the feedback. I think the main difference between the products you mention and BackendLab is that BackendLab will soon offer the ability to export your code (this isnt yet mentioned on the landing page). So you can build your prototype, export it then build it out. The export would initially be done as a Django project, as that's how the BackendLab api is built

This sounds great. I decided to try it out after reading about this export feature.

Im considering signing up and trying out your service. What is holding me back are all questions I was hoping to have answered on your website. The main questions I have are important implementation details that I think customers a going to care about. What language, framework, dependencies, are handling requests? What is the performance like? How do I scale?

Also, slow down if youre shipping typos on your landing page: revolutonizes

On the other hand, as the saying goes, if you are not embarrassed by your first version, you waited too long to ship. ;)

Exactly :) I'm aware that there's a long way to go yet, but ultimately its still just a beta

Thanks, this is good feedback. The landing page was deployed in a hurry and will be updated soon, both to fix typos and to add some more technical information about how it's built.

For the record, it's all hosted in AWS EC2/RDS, and the backend is built in Django. The front end is nuxtjs

Thank you, this is a key consideration. May I suggest language like "Powered by Python and Postgres"? If youre using the ORM, or wrappers or middleware on top of django rest framework, then I see your target market being front end devs only who need a scaffolded back end. I say this because the service doesn't remove much work for users with API design and implementation experience.

That's a very good idea, I'll do that

It's a good point. Running your copy through Grammarly will help.

We have a similar concept but with frontend and permission as well. The backend is automatically created based on what is defined in the frontend. Makes it more of a business tool than a developer tool. For developers, we have CRUD APIs. APIs make it useful to collect data from different hardware devices.

Here is the link if you want to explore: https://www.clappia.com

Some examples on the website would be helpful to understand the experience. How do I interact with the tool, what does the output look like, how do I use the back end?

On a similar note, what would be a good option for building front-ends without much coding? Maybe something that could export a React project, for quick prototyping.

What would this look like? I understand that there is a specific pattern repeated over and over again with models authentication users and CRUD for a simple backend but even a simple front end seems much harder to implement in a generic way. Maybe some form of templating?

I would think something like Rails scaffolds, only it generates React code. Heck, you could probably extend Rails scaffolds (or other frameworks that provide similar utility) to accomplish this.

I'm working on exactly this. I don't want to self promote too much in this thread, but feel free to shoot me an email at gabe.ragland@gmail.com and I'll get you in the private beta.

There's a long history and many tools for that. You can search for UI generator, UI Builder, CRUD or RAD (rapid application development).

For React: Structor, React Studio, Pagedraw.

I like the look of this. A couple of questions I can't see answers to on the landing page... - Is it possible to change the schema after creation? Or is this a "get it right first time" deal? - Is validation strictly not-null/range-check? Or can you add more complex validation rules?

Hi, thanks for looking. You can edit your data models at any time - fields can be added/removed/renamed as necessary. There is some validation already - string stra fields must be strings, number fields must be numbers, etc and you can make fields as required (meaning that it a value must be supplied), unique (meaning that two instances can't have the same value for a particular field) and also set default values. Additionally, foreign key and many to many fields are validated to ensure the related objects actually exist.

I'll also be adding additional validation (e.g. regex, max/min numbers) in the very near future

Small note:

http://www.backendlab.app fails to load. https://backendlab.app gives a security error

Thanks, but I'm not seeing the first issue - it automatically redirects me to https. What browser are you using?

Saw it on Firefox on Fedora. Today at work it works with Chrome on MacOS.

Looks like a very interesting alternative to Knack (https://knack.com), but without the frontend building aspect.

One question I have with a project like this is: suppose I used it, and got traction. What is the story if I then realize that I need to migrate to a more traditional architecture?

Hi, I've replied to this in another comment - BackendLab will soon allow you to export your app, as a Django project (the BackendLab API is built in django). That way, you can rapidly prototype using the interface, and then export and build it out further outside of the environment.

The fact that the landing page doesn't yet mention this is a big miss - will try and get that updated soon

Take the code they give you and change it?

I think that's the ideal story: the system generates an intermediate representation that can be interpreted or compiled to a programming language. There are bindings to common languages/runtimes, and you can take your system and leave at any point.

In practice, lock-in seems like a real issue.

They dont give you code. You get an api with a postgres db on Amazon rds that you manage through forms.

If maybe you could allow few lines of code execution that can work for many devs, then they could recommend it to beginners after community emerges :/

Eventually I'll be adding the ability to add your own endpoints to the auto generate ones, which would initially give you the ability to implement various additional functions, e.g. sending emails, integrations with other services, etc. This could very easily be expanded to allow users to call functions they have written themselves. However, I've been holding off on this one as BackendLab is primarily a no-code solution. If enough people are keen, I will implement it

this is awesome. looking forward to a couple years a of fun and easy prototyping before this inevitably gets bought and shut down by some giant

I reviewed the site. Really nice work. You can sell additional features like the live monitoring tool. Heroku add-ons style. Good luck.

This is a pretty great idea! I've always wondered where "MS Access for the web" is, ie RAD DB development.

add roles in addition to groups.

these are different logical segmentations that are nearly always conflated together in most "quick crud" systems, and never map to reality on things much more complex than "todo list" API stuff. Or, perhaps more specifically, in anything that might end up being considered a 'line of business' app.

There's no email verification, I have been able to register using plop@plop.com. Feel free to delete this account.

Yeah i know, its on my to do list...

Hi Chris,

Great Idea I guess some work needs to be done regarding building a frontend for your idea

Best, Victor

> Stop writing code. Start deploying apps

I find this is close to sarcasm.

Trying to remove code from creating apps seems like an oxymoron to me too. Eventually you need constructs which code can efficiently provide.

    if, for, while...

Sir, please try our product Clappia (www.clappia.com). You are right that you need IF, ELSE etc. You can use the formula syntax of Excel sheet and you can still manage without learning any new coding script. That's what we do.

Hey Chris,

Congratulation for launching your beta!

This seems quite similar to what we do here: https://cloudbackend.appdrag.com/

But it's look like your auto generated API support only CRUD operations, this can be very limited even for common use cases like: "Create a new order" make some business validations, store in the db, generate a pdf, send an email ...

Disclaimer: I'm the CTO of AppDrag

Reply on news.ycombinator.com