I was asked an interesting question on here recently via PM. That is, “As someone capable of doing their own coding, why do I use widgets?” Several reasons actually, but before getting into that, let me explain what makes Elfsight widgets unique and why I often choose them for my own or client applications.
Perhaps you’ve noticed the name disparity in the install code, “https://elfsightcdn.com/platform.js” using elfsightcdn.com instead of just elfsight.com? That’s because Elfsight widgets are delivered via a cloud-based Content Delivery Network (CDN). This does two things for you.
Your site will not load slower as a result of adding the widget, since it will be served from the Elfsight cloud vs your own server.
Some visitors may actually benefit by your page loading faster than if the widget had come from your site. This is because cloud CDNs use a distribution technique known as geotargeting, which serves content from the datacenter nearest to the requester.
Other Reasons for Using Widgets vs DIY:
It’s easier and faster to implement than doing it yourself. If you do this for a living, you quickly learn that time is money!
It’s been thoroughly tested before release. If there turns out to be a bug or oversight, it’s quickly fixed.
There is a vast pool of resources to draw from. Do you have an idea, question, or suggestion for improvement? Between the support crew, developers, and fellow community members, you should have no issue getting an answer or brainstorming an idea.
Case in Point…
I recently had an immediate need for adding tabs to a page on my own site and inquired about a Tabs widget ( It'll be great to have a widget for creating tabs (Tabs widgets) - #31 by Bob_Wilson ). It’s likely to be some time before (or if) it happens, so I created my own that are both accessible and responsive. But, it came at the cost of considerable code added to the site. If a Tabs widget is released in the future, I’ll switch to it in a heartbeat!
A skilled coder certainly wouldn’t wait around for a widget just to handle tabs.
A layperson, however, would—since they have no other alternative… other than learning to code themselves.
But this raises a question for me: would a coder like you venture into more complex widget alternatives? I’m talking about examples like reviews, social media feeds, AI chatbots, etc.—basically, tools that require integration with external platforms and might even involve licensing fees to gain access.
I suspect that’s a scenario where even a coder would choose to wait it out… wouldn’t they? I tend to view this arrangement as a sort of “group buy” or wholesale access model; that’s where the cost-benefit ratio for the individual user becomes most compelling. After all, I’m a solo operator—not a full-blown web services agency. That’s precisely why I appreciate being able to “piggyback” on Elfsight’s infrastructure. It’s a win-win-win situation.
Good question. Whether to wait or develop it ourselves I think would depend on the nature of the request. External platform integration or licensing fees would not likely be deciding factors, as that happens often.
Elfsight and individual developers have fundamentally different endgames. Elfsight designs their products to fit a variety of circumstances and needs. For example, there have been recent discussions here demonstrating how people are using the AI Chatbot and Form Builder widgets in various clever ways. This happens because Elfsight builds-in the flexibility and customization capability. We design to meet or exceed customer spec… period. When a customer comes to us requesting something for which there is no reusable code and development effort exceeds what we could reasonably pass on to the them, we would prefer to wait. If not providing the feature at launch was a deal-breaker for the customer and we couldn’t convince them that deferring the feature would be beneficial, we would probably refer them to a friendly competitor. We would lose the sale, but there is something to be said for the goodwill and trust it would generate. I think a combination of custom code and customizable widgets is the sweet spot for most people. DIY website builders augmented by widgets are also a popular choice.
I especially liked your point about combining custom coding with customizable widgets. You really don’t have to choose one over another, as widgets can fit naturally into a custom workflow and take a lot of work off your plate.
And since widgets keep evolving with new features and improvements over time, that approach can save quite an amount of effort as well
Thanks for the information, really helpful. I am a marketer and can’t code so using Elfsight for years. But since AI came along I am learning vibe coding and was able to solve quite some things without knowing coding. Could ir be that vibe coding becomes another alternative to custom coding? Meaning you can custom code but with the speed of using a ready widget?
That seems like a reasonable approach. The thing to be careful with when using AI in this way is that you must be very specific with your request or you may get an incorrect answer. For example, while working on a site recently, I needed to do a server-side redirect for an individual page. This is normally a mundane task, but it wasn’t working for me. After getting an obviously wrong answer the first time, I re-asked the question, clarifying that the ‘from’ and ‘to’ locations were on the same domain and that the URI contained encoded spaces (%20). That got me closer, but gave incorrect escaping rules. Finally, I realized that it wanted to know what type server was in use and what version server software.
You see the problem… the answer to some questions requires you to provide information that you didn’t even know was relevant. But when it works, it can work very well.
Thanks Bob :-). Yes, you are totally right. I have/had the same issue creating content for my clients. For example writing a blog in the voice of my customer. I work with AI now for over 2 years and as you said, this is the key “must be very specific with your request”. I actually now have two otherAI’s where I double check the first AI’s response, which helped a lot. But again, you the human are still in the loop. Lets 's see where this goes with the AI …
At least for now, it still seems difficult to fully delegate everything to AI (whether it’s vibe coding or other AI-assisted tasks) without double-checking the output. Which means that some level of expertise is required so you’re able to evaluate and validate what AI suggests.
@Ralf_Gerhardt you mentioned you solved some tasks with the help of AI. Would be really interesting to hear what kinds of things you’ve managed to delegate or build so far
