But it's not a great alternative to a DB. It's really expensive per seat (twice the cost of Google Apps and you only get one app). The row limits on tables are too low. And I'd really like native webhook support to ingest row updates.
I'd be very interested an Airtable clone that I can run locally on-top of Postgres or Mongo or something like that. My main use-case would be to replace expensive-to-build internal CRUD apps that really should just be spreadsheets but require bespoke integrations with other internal systems.
I've started working on a similar platform where the goal is to make building CRUD apps wildly simple. The plan going forward will be to layer on more sophistication after the core foundation is extremely solid.
The platform uses Mongo for the dynamic data tables which is a game changer for tools like this. Postgres jsonb data would also work, but Mongo just feels more natural to me.
I've named the platform Webase. Check it out here: https://www.webase.com
Note: this is very basic currently but consistently getting better!
> You have not created any apps yet.
with no way to create an app that I can see
> Granular permissions not only at row level but field level too. You can share parts of data with others for editing or have it view-only with ability for commenting and accepting suggestions
In my org (and I suspect many), many of the Sales/Operations people maintain Google Sheets with links into our internal apps. Generally it's TODO lists or other little organizational/workflow type stuff
What I think makes Airtable so powerful is that foreign-keys are first class. You can create a table and include references to rows from other tables. This allows the end user to create tables to support their workflows directly in the same tool they use to manage the data, and means I don't have to build that same functionality into the CRUD app (and until you're deep in a workflow it's hard to get it perfect, so there's always back-and-forth).
A good permission system would allow me to create a set of "core tables" that I tightly control the schema of, but allow others to still create tables which reference this one.
Do you have some examples?
We're building a sort of highly linkable (cell level) spreadsheet-type widget builder that is supposed to play nicely with existing workflows (instead of trying to force people to build out entire workflows like many of these new breed of "CRUD-replacer" low-code/no-code apps seem to want us to do). Might interest you!
We have a CRUD app that acts as a CMS with all our product listings in it. There is a team of people who are tasked with curating and maintaining these product listings. So you can go to crud.com/admin/product/<id> and modify the listing, which needs to happen every now and then.
The team of people who maintain these product listings need to coordinate their efforts and so when they need to modify some set of the products they create a spreadsheet with (product_id, link_to_crud.com, Person who is responsible, status). Maybe there's some approval process or multiple statuses or something like that that makes this all a little more custom/complex. Maybe they need to update a few things across services at the same-ish time so there might be a few links per row but it's going to be relatively bounded by what a human can handle.
If our data lived in Airtable, it'd be really easy for this team to make their tables/spreadsheets with direct relations to our core tables (which replace the CRUD app in this case). That team understands their workflows better than I ever will as an engineer. Giving them the power to create their own workflow tools is way better than trying to have engineering build it for them, and that's going to be a huge selling point for me.
I don't think we'd need cell-level links but rather row-level. Right now they use Google sheets, but a more domain-specific tool would be better.
What I'm working on though is more about taking the core of existing spreadsheets and building powerful integrations on top of it. Like in the sibling comments, building a webapp, or something else. I really think this has huge potential.
JS is nice because all you need to write it is a text editor and a web browser, both of which just about all computers have. I had an entry-level job that involved a computer once and automated a lot of the painful stuff with JS - I bet that if more people knew JS, more people would do the same.
I don't know if that'd be a good business proposition, but I think it could be a good on-ramp.
source: I made one myself (and know the founders of others)
AirTable has made huge inroads here. I'm also building something in this area (https://lightsheets.app/), turning slightly back towards spreadsheets instead of databases but then building enhancements and modern integrations on top. I think there's still lots of potential in this area, despite spreadsheets being 50 years old.