Hacker News Reader

Comment by rubenfiszel | Hacker News Reader

Congrats on the project and company, looks really cool! Funnily enough I wanted to write the same post this week. I got lucky enough to get into this current batch of YC, but in the end you ended up in a great spot if you were able to raise directly 1.6M. Our tools have so many similarities, Windmill is also AGPLv3 (https://github.com/windmill-labs/windmill) and also self-hostable, and also about building custom internal tools. We focus a lot less on UI but on the run custom logic part as flows of Deno or Python (and we actually just released an open spec about it we call OpenFlow: https://github.com/windmill-labs/windmill/blob/main/openflow...). So in that sense we could be greatly complementary. If you would be interested to chat: ruben@windmill.dev.

View on HN

It is just my opinion that market opportunity for a tool like this is actually quite small. There are two constant struggles that I can of that you might be facing. People find the product great but don't know what do build with it. It is hard to be in front of the customers when the painpoint is there, and when the painpoint is there they might be solved it is some form. Secondly, developers will trivialize the product. Developers might say things like if I have to run scripts I would rather make it part of the service or it is just a UI to run my scripts, I would rather run it from my machine. Also, you can't really sell to larger orgs, tools like this don't have place and will be competing with retool and others. Don't want to discourage you but it is going to be a uphill battle. With works like a temporary replacement for CRM, backoffice tool for a very short window in small orgs. When they grow, they will either adopt point solutions or retool. Just my opinion.

> People find the product great but don't know what do build with it.

On https://hub.windmill.dev/ we are building tons of flows that you can import in one-click (also available from the product itself) so you get ideas of what to build with it.

> Secondly, developers will trivialize the product. Developers might say things like if I have to run scripts I would rather make it part of the service or it is just a UI to run my scripts, I would rather run it from my machine.

That is one of they key audience that I'd like to convince. I am a very skeptical and high-requirement developer myself and I would have loved to have had a tool like this. What's not to like, the performance overhead is very small (thanks to rust and v8 integration), and it saves you a TON of time to compose scripts the right way. You can self-host it so you can just 'run it from your machine'. You can make it part of your service, it exposes HTTP endpoints for both triggering the script/flow or just waiting for the script to finish and return the result, like AWS Lambda. Having your own self-hosted AWS Lambda does not sound nice ?

> Also, you can't really sell to larger orgs, tools like this don't have place and will be competing with retool and others.

What's your argument there ? You cannot ever sell to a larger org because there is a product that already exist and that already sell to larger orgs ? Being open-source and having openspec is imho a big competitive advantage for selling to large orgs.

I just took a quick look at hub. I saw the example for datadog and pipedrive. Datadog has a script to post metrics. I cannot think of a situation where I would use a script like that, you would rather want it to be from the service itself or the agent. To me it sounds like an anti pattern too. But that's just me. Secondly, for pipedrive i see scripts add user etc. I think in enterprise setting i would use a platform like better cloud that specializes on saas user provisioning. The thing that you are missing and lot of platform in this space is that if you market as it as a devtool then it is not good enough, dev workflow looks very different and you are force fitting this. What has a more cohesion? Adding a script in git, clone and executing it via your machine/teleport/strongdm etc or your service, that makes debugging through stuff more difficult, completely unnatural experience. I think self hosting AWS like lambda is a bad idea because there is huge operational overhead at scale. If you position yourself as CRM alternative, then it makes sense for smaller orgs, but then my argument holds because the moment they become bigger they will transition out of upper platform. Openspec is actually not a differentiator here, there is core problem, when the org mature they invest in point solutions to solve particular issues. I mean well for windmill, may be you have a better insight that I dont have. I have been tracking this space for a while and these are just opinions in the end and might be wrong.

Had seen Windmill a while ago, nice work and happy to chat, let me send you an email to get connected.

Reply on news.ycombinator.com