Dropserver Progress - March 2026

This is the progress report for Dropserver for March 2026. The previous report is here.

This is going to be a short update because I got very distracted with a different project midway through the month.

Continuing Work On Removing DropIDs

See the previous few updates for more info on this.

I continued to grind away, making progress and pretty close to having something releasable.

However before mid-March I got sidetracked with a new project, so progress on Dropserver stopped for now. This new side-side-project circles back to Dropserver and Dropserver apps eventually, but I am pausing work on the core of Dropserver for a little bit.

Then came an unhelpful item…

Deno Questions

Followers of the Deno project were gripped by a sudden wave of personnel leaving the company. No formal comment from Deno or its creator Ryan so we just don’t know what’s up. But it’s safe to assume it’s not a good thing.

Deno is what made Dropserver possible so its demise would be a real blow.

Would the disappearance or abandonment of Deno be fatal to Dropserver? I can say it would be bad, but there would be options to move forwards.

Node sandboxing

Node has added sandboxing (ok permissions) to its runtime. I don’t know how solid it is, and since they are bolting this onto an old codebase there are bound to be leaks. There are a few gotchas (see the bottom of the link above) but it’s there. It’s a thing. I could just switch everything over to Node.

WASM

I have been tickled by WASM. The efficiency of a WASM-based solution is enticing. But what you gain in efficiency you pay for with gnarly developer experience. There are so many acronyms floating around WASM you’d think it’s a NASA mission. The learning curve is steep.

I love that with Deno-powered Dropserver you can write a short JS file and you have an app. Done.

There is also no standard WASM precursor language. So different devs will write in different languages, diluting the ecosystem and hurting my ability to help them out. The WASM ecosystem is still evolving significantly. As far as I know you still can’t link a bit of WASM written in Rust with some WASM written in python for example. Lots of challenges there.

Deno is still here for now. Onwards.

Dropserver and Vibe Coded Apps

This progress report sounds pretty bad so far, sorry. But here’s why I’m still fired up about the project, even if I am not in the codebase right now.

The other day I had a home computing fiasco on my hands. I’ll spare you the details, but let’s just say I had to move some files around different computers and the usual way I would do this wasn’t available on all computers (USB drive, email, Tailscale).

A Dropserver app could help: I just needed an HTML form to upload a file, and list the uploaded files with links that cause them to get downloaded. Super easy, and I could do this myself in a small amount of time.

But rather than code it myself (I was frantically trying to get back to normalcy) I asked Claude to build it. In part to see if it could do it, and also hoping I could get away with not having to actually write it myself.

Amazingly it built it in about 90 seconds. I just had to fix a couple of Typescript errors and that was it.

I had no qualms uploading it to my home ds-host instance knowing that Dropserver would sandbox it. The app could not make any requests to the outside from either the server side or the browser and can only read its own files.

And that’s why I think Dropserver is very relevant today. I see people generating apps for small personal itches all the time. Dropserver makes it possible to build a small multi-user thing and run it in a sandbox. If your LLM produced unsavory code that tries to send your data to some outside site, it won’t be able to. Even a memory-leaking mess of an app would be fine because ds-host automatically kills apps when they are idle.

When I started Dropserver a big worry was that its inability to run standard Linux apps meant there would be no “app store” at the outset. It’s why it was so important for Dropserver apps to be easy to write. I never could have predicted the world we live in, but it turns out apps are now even easier to produce. And that puts this project in a good place.

Til Next Time

We live in challenging times. There is pain and uncertainty everywhere. Still, I’ll see you next month.