It seems odd that there is nothing like Access (or the other couple if similar DB tools that were around in the heyday of Access and before), but I think it is because the demand is relatively small. I've seen several projects start in that direction then pivot elsewhere or just die an unsupported death.
I think the problem is that demand/interest is not sufficient to keep a self-hosted project going, nor monetisable enough for a hosted one. People wanting to self-host end up going with something more specific, possibly self-made, for their needs, rather than a generic solution, and a hosted solution has a couple of significant costs to cover:
1. Resource use when people load a large amount of data then run under-optimised queries on it (or impossible to optimise, if they've chosen a bad structure for what they want out of the data). This can be mitigated by throttling individual users' IO/memory/CPU use but then the product gets a reputation for being slow.
2. The support that many people will expect (especially if they are paying, but even if they are not) which could consume a lot of time. A project that is very lucky might end up with a community that takes on a good amount of this load, but you can't bank on being that lucky.
3. Resource to keep available all the hardly used, or even never used, projects that will sit around if the service is free. Mitigating this with cold storage will help, but as with throttling active use this will make the service appear slow generally (people will remember the tens-of-seconds startup time more than they will notice subsequent actions being more than fast enough).
Getting people to pay will be an uphill struggle, and money from advertising is unlikely to cover the above, especially with many people like me blocking commercial stalking which also blocks a lot of advertising.
There are many commercial services and tools. Like Notion and Airtable and all their clones. For more advanced users and usecases there are those like Metabase, Retool and all their clones. But they are more focused on specific domains. And today it's quite easy to just barf up some CRUD-interface with webstack, especially now that AI is good enough for simple stuff.
I think the problem is that demand/interest is not sufficient to keep a self-hosted project going, nor monetisable enough for a hosted one. People wanting to self-host end up going with something more specific, possibly self-made, for their needs, rather than a generic solution, and a hosted solution has a couple of significant costs to cover:
1. Resource use when people load a large amount of data then run under-optimised queries on it (or impossible to optimise, if they've chosen a bad structure for what they want out of the data). This can be mitigated by throttling individual users' IO/memory/CPU use but then the product gets a reputation for being slow.
2. The support that many people will expect (especially if they are paying, but even if they are not) which could consume a lot of time. A project that is very lucky might end up with a community that takes on a good amount of this load, but you can't bank on being that lucky.
3. Resource to keep available all the hardly used, or even never used, projects that will sit around if the service is free. Mitigating this with cold storage will help, but as with throttling active use this will make the service appear slow generally (people will remember the tens-of-seconds startup time more than they will notice subsequent actions being more than fast enough).
Getting people to pay will be an uphill struggle, and money from advertising is unlikely to cover the above, especially with many people like me blocking commercial stalking which also blocks a lot of advertising.