Transcription
I have tried more than fifty no-code tools. And in this video, I will tell you about the best of them. That is, these are services where you can create a mobile application, a website, maybe a game yourself, and do it all without developers, without code, and in the shortest possible time. Each of these services has its own nuances, so watch the video to the end so as not to miss anything. My name is Ignat. Let's go.
Perhaps you have heard that artificial intelligence is replacing programmers. In fact, this is no-code, when it is enough to describe your idea in understandable Russian and get an implementation. That is, artificial intelligence itself will write the code for your application, and you just have to steer this process. And now web-coding is the new oil. This technology is used by aspiring entrepreneurs and startup founders who want to get a product as quickly and cheaply as possible and start earning. Marketers use it to quickly build a website, get clients, and understand the effectiveness of their chosen advertising. Managers use it to automate some routine processes in the company and free up more time for strategy. And, of course, artificial intelligence and development through it are used by programmers themselves to simply work more efficiently. But before we begin, a minute of self-promotion. Artificial intelligence works wonders and can create a "wow" effect when you enter a couple of sentences of text and immediately get a meaningful result. That is, you get an application that generally works. But behind this facade of simplicity lies many pitfalls. These are logical errors, security issues, limited scalability, and difficulties with refining and supporting the product after launch. And if you want to develop a full-fledged application without limitations and other problems, then follow the link in the description. We have affordable prices and reasonable implementation times. Well, if you are interested in learning to create applications yourself and understand the process better, then check out the Telegram channel. There I show how I create applications and write a lot about the technical side of this process. The link will also be in the description.
So, there are now a huge number of services that offer no-code development. Just so as not to waste your time, I will tell you about only four that have their own clear, distinct niche and solve tasks within that niche better than others. So, the first service is for those who do not want to look under the hood and do not want to understand how this thing works. That is, you don't know what a database, API, frontend, backend are. You just, for some reason, need to get some kind of website or application or implementation for some idea of yours. And in that case, I recommend looking at the service B44. B44 is a product of the company VIX, which is vix.com, a website builder. Well, now, the company VIX, apparently, is trying to keep up with the times and is entering the AI development market with its constructor, essentially the same boxed solution as for websites, aimed at the same audience of non-technical people, completely unskilled, who just need to take and do it. And the simpler they can take and do it, the better. Please, B44. Even your grandmother can create something meaningful and working. That is, there is a large input field. You type your fantasy there, what you want to get. You press a button, the system thinks a little and gives you what you wanted, what you requested. And it does it quite well, it must be said. That is, you have no ability to edit the code, you cannot choose a stack, you cannot design databases. But if all these words are gibberish to you, a great choice. Take BAS4. Indeed, you will be able to do something meaningful. That is, they have quite well refined their algorithms, their AI, their technology to truly cover a fairly large range of tasks in this way. What are the limitations of this service? Well, actually, one big one. That is, all the work in Bate 44 for you, as a developer or entrepreneur, or anyone else, is just a big black box. You send a request there, something happens there, and you get some result. You don't see the code that is written. You don't see the integrations that are chosen, how security is configured, access to the database. You don't know anything. You just write a letter, I don't know, to Hogwarts, and then oops, magic, and something appears. This magic doesn't last long. That is, if you try to do something more or less complex, if you don't send one letter to Hogwarts, but iteratively send and try to develop and refine your service, send five letters, ten letters, send some clarifications, and so on, it will work worse and worse and worse. It will mess up, it will break what seemed to work before, and so on. That is, you simply cannot create anything really large, complex, serious, or loaded out of the box without diving in. Again, from this simplicity, from this format that is at the very foundation of the product's idea, that a non-technical person can do and develop everything, the following disadvantages also arise: you are completely dependent on VIX's infrastructure. That is, your application runs on VIX servers, you pay a subscription, you cannot leave them. If they are blocked in Russia again, you can't do anything about it. Your application is completely dependent on them, on their infrastructure, and is designed and operates only on their infrastructure. You cannot take it away from there. Well, another minor limitation is that on Base4 you can essentially only create a website, i.e., only a web application. You cannot create a mobile application or a game there. The infrastructure is simply not for that. Somewhere, possibly in some advertising articles, perhaps even BS44 itself will write that you can create a mobile application with them. However, this is a clever marketing ploy. It is not a full-fledged mobile application, it is a so-called webview. That is, an application that is essentially a very, very stripped-down browser. That is, a browser application that displays only your website. That's all, you can't go anywhere else. There are no bookmarks, nothing. Just your website is displayed in the browser. That's it. But your website is still your website. Everything you programmed there is a website. You can wrap it in ten such webviews, but it will still be a website. The problem with such web technology is that Google Play App Store moderation is very reluctant to approve it. By reluctant, I mean that in 95% of cases, it is rejected. That is, if your entire application from start to finish is a webview, it is not published now. Therefore, when choosing Base 44, you must understand that you only get a website. That's it. If you want to do something more or less serious with a view to scaling, even just for hundreds of users, with the expectation that you will refine and develop it further, then I wouldn't get involved here. You will have to develop and refine everything on the fly, without any control points, with huge risks of breaking something in the process and incurring significant reputational and financial costs.
And briefly, let's discuss payment. Look, like all the services I will mention in this video, BAS44 earns through subscriptions. Like all services in this video, the basic subscription costs $20. Like all services in this video, you cannot pay for this subscription with Russian cards. But over the years of being in such an unpleasant situation, of course, a lot of different solutions and workarounds have appeared, which, among other things, are suitable for Russian citizens who do not want to invest additional money in obtaining foreign cards or anything else. To get out of this situation, I would recommend getting a BYBIT crypto card. An important point is that this is not an advertisement. I use the card myself. I know the services where it works and where it doesn't. Therefore, for these specific tasks, I can recommend it. What's the point? Bybit is the largest crypto exchange. Everything is legal, everything is official. An exchange where you can still register using Russian documents. There are no prohibitions. That is, you simply register, and the card is issued for free. And the card simply spends money from your Bybit balance. Again, everything is one-to-one, without any commissions. The card can be linked to Google Pay, Apple Pay. The card can be used to pay for all subscriptions, including those I have outlined in this video. A detailed instruction on how to register on Bybit, how to top up your account, how to get a card, will also be attached to the video description. Please, everything is open, everything is free.
The second service in my top is a bridge between BAS44, when you just create and forget, and more or less serious, complex development through AI, when you can really do anything. In fact, this golden mean is called Lavable. And when you go to the site, you might ask yourself: what's the difference from B44? There's the same input field in the center, and it also says: "Tell us about your idea, and we'll also make an app for you after some time." The trick is that, unlike B44, on Lavable you can take this application, refine it, and get not just a one-time creation, but perhaps even something to develop and scale in the future. The key difference of Lava was that it doesn't trap you within the VIX ecosystem, as is the case with B44. Lavable uses the best solutions currently available on the market and allows you to work with these solutions independently. For example, as a database, Lavable strongly recommends using Subbase. This is an open-source platform, a backend as a service. That is, roughly speaking, everything is collected together: the database, file storage, the entire authorization and authentication system, and automatic scaling. By default, this thing, this SUBAS, runs on Amazon servers. But with very little effort, you can take this Supay, download it, and install it on a Russian server, for example, to comply with Federal Law 152 "On Personal Data" for example. That is, Lavable in this regard gives you many, many more tools to do what you want. Yes, everything is super simple. Yes, you also enter a request and get some solution, but the trick with Lavable was that you can go through some path by developing only through AI, i.e., by simply writing your requests. But at a certain point, when you hit a wall and when the AI's capabilities are no longer enough, when it starts hallucinating, when it glitches, you can take this thing, download it, hire a programmer for some money to figure it out and make it work. But initially, the technologies, namely Typescript for the front and SUI for the backend, are chosen in such a way that you can generally get a good working service that will withstand loads, that will scale to, say, tens, hundreds of thousands of users, and so on. But this is no longer within the Lavable ecosystem itself, but after you take this code and finish it. The second limitation is that Lavable works with the backend quite limitedly. That is, some basic things, like, say, we have an application, the user enters an order, this order is saved in the database, or they write a message, and the message is saved in the database. This is called CRUD operations in programming. That is, a very basic history. Simply create, modify, delete, and something else. All this is done out of the box precisely due to this integration with Supabris. But if you suddenly need to write some custom logic that is not available out of the box in Supab, Lavable will have difficulties with this. For example, you want to implement payments in the application, and you want to accept payments from Russian cards. To do this, you need to connect a Russian acquiring system, for example, YooKassa. To connect YooKassa, you need to write two or three simple functions with code on your server. And Lavable will not do this for you. You cannot connect this thing. That is, it's more for the frontend. You can purely create beautiful applications, good-looking ones, set up the interface, and make the buttons clickable, ready for production. Or either not, or you will have to download and finish it. And, of course, the caveat is that this situation, for example, with connecting Russian payments, is difficult to implement, and for BAS44 it is impossible to implement at all. That is, all the technical limitations that apply to Lavable, of course, apply to B44 to the same, if not greater, extent. And the third limitation of Lavable, I would still note that here you need, well, to understand a little what you are doing. That is, without any knowledge, from scratch, it will probably be quite difficult. You will have to at least read what a database is, register in several services at a minimum, get keys, insert all this into the right place, configure it, and so on. And even within this short speech of mine about Lavable, you probably see how many times I use quite technical, specialized words, which are actually difficult to avoid, because, well, what can you do. For a complete novice, who just came in and knows nothing, but wants to, it will probably be difficult. Who is Lavable suitable for and why is it better to use it? Look, Lavable is a great thing to quickly create a sketch of your application, to create a draft. Then you will look at this sketch and make a decision for yourself. If you like this sketch, you can find tools to color it and turn it into a full-fledged picture. If you don't like it, throw it away and forget about it. That is, this is the key, this is a fork in the road, that GLAV was super for creating an MVP. But unlike base4, which you won't have to throw in the trash after you use it, but which you can take and simply develop further, I don't know, independently or by asking some developers to finish everything with code, build it up, and so on. That is, the foundation that lavable provides allows you to do this. Also, if you are a developer who creates applications, for example, as a freelancer, then Lavable can be a very cool catapult for you, an accelerator, that you've put together 80%, 90% of some functionality, somehow assembled it, more or less something is clickable, there are a lot of bugs. You download all this, set it up locally, and start finishing and fixing it. And for me, Lava is also a good tool to see what AI development is, how it looks, what its limitations are, what bonuses it has, what works well, what doesn't. That is, instead of listening to various bloggers and watching reviews, you can go to the site for free, you don't need to install anything, it even has some free credits on Laavable, so you don't even need to pay for a subscription, take it and try it out and just see if you like this format of work, how much faster it does what you want, or maybe it slows you down. That's also possible. And perhaps you've noticed that we're going a bit by increasing difficulty from something very basic to something a bit more complex through lavable. And the logical conclusion is what tools to use to do absolutely everything you want, and create such "dirt and filth" that previously required a degree in Computer Science, and now it costs literally nothing and I don't know, one interesting evening with a cup of tea.
I want to take a short break for design. In fact, when we talk about creating an application and how it is done, when the code is generated, it is actually very good to draw the application clearly beforehand, because the user doesn't care what you develop the application on, they will look at the storefront, at the cover, how much they like this design. What to do if you don't want to deal with designers, if you don't want to hire them and pay, there is a service UX, which allows you to generate logic, i.e., UX, where the buttons are, where the input fields are. And it also allows you to color everything and generate a fairly, well, understandable, decent design. First, you write a technical specification, i.e., what you want to see on this screen. You can write it quite basically, for example, a technical specification. That is, I want to make, I don't know, registration, within which I want to ask the user about this, this, this, this. Just list it. You can put your description into ChatGPT and ask it to specifically write prompts for UX file for generating such an interface. The key thing that I like most about this service is that it allows you to generate not just one isolated screen, but a so-called flow, i.e., a chain of screens, a chain of actions, for example, registration. That is, as a rule, the user enters, chooses to register or log in on the main screen. Then they enter their phone number, then on the next screen they enter a code, then on the next screen they enter some information about themselves, then, I don't know, they choose, I don't know, a delivery address or something else. That is, it's a chain of screens. And UX Pilot allows you to generate all these screens at once. That is, in UX Pilot, you can generate screens, essentially, at a fairly high level, but still, it generally helps a lot for the entire application. That is, instead of coming up with it on the fly, how everything will look in your application, while doing this development through Lavable or through Base44, well, you initially wrote everything down on paper, got a cool screen map, everything is clear, everything is visible, everything is great, everything suits you, and then you go to implement it. The question is, what to do with this screen map? UX Pilot allows you to import the screens created in it into the FigM application. This is, well, the industry standard for designers. Everyone works in it. It is completely free. That is, you import it into Figma, and in Figma there are plugins that allow you to convert a Figma screen into code written in various programming languages, for example, in React or Flutter. So, you create the design in VX Pilot, convert it to Figma, and from Figma you get the code in your desired programming language, for example, in TypeScript, which you then send to the tool you are writing on, I don't know, again, the same Lavable. You send it, please, here, I've sent you the design of the main screen. And it says: "Okay, good, I understand, this is the design of the main screen." And then, the Lavable itself. That is, it's enough to just send some basic screens. Then it will also do something for you in roughly the same style. That is, your fonts will be set, your colors will be set, the general visual style, the general mood of the project, will be set by this UX pilot, what you import. And as for some micro-moments, micro-screens, the development environment itself, so to speak, will offer them to you. On UX Pilot, you won't be able to create a really serious and detailed design for a project. For example, we do custom designs when we work. That is, there are several hundred screens for an application to highlight all possible statuses, all possible screen states, and so on. That is, it does not generate this. With its help, you can just sketch, so to speak, the skeleton, a few key screens that will set the overall tone, style, mood of the application, and from which you can then proceed with development.
And now we have finally reached the main tools with which you can essentially not limit yourself in anything, which allow you to create complex, commercially viable products, products that can be further developed, scaled, launched, I don't know, into new markets, onto new technologies, which are tools that allow you to create websites, and mobile applications, well, normal mobile applications without any "ifs" and "buts," allow you to create games, virtual reality applications, for smartwatches, Telegram bots, parsers, anything else. And it is precisely these tools that embody the idea that AI is replacing programmers. The AI development that is changing the industry and due to which a huge number of programmers are losing their jobs looks different. These are tools that integrate AI very competently and conveniently into the usual application development process. Therefore, to use this service, we don't need anything, i.e., it's not some kind of ready-made product, like, go to the website, register, link your bank card, and you can use it. No way. This is the approach itself, which I strongly recommend everyone to try themselves to see that it really evokes, let's say, effects and emotions, especially if you know what it's all about. There are several applications on the market that allow you to do this kind of "dirt," to create anything you want just from your head, and with quite good quality. The most famous is Cursor, but I would not advise you to go in that direction, simply because it will be expensive. Cursor has gotten a bit greedy. In my work lately, I have been using Google AntiGravity. Simply because? Because it is much cheaper and more comfortable. That is, Google now has a comprehensive, pleasant subscription, Google AI Pro, which includes good request limits for AntiGravity. Plus, it includes a lot of other perks. This is video generation, this is, I don't know, image generation, and Gemini, and 2TB of hard drive space, and so on. That is, it's a really good subscription. Google is clearly trying to monopolize the market, offering a lot of valuable and tasty things to attract people. Therefore, at the moment, it will be most comfortable in terms of cheapness to work through Google. It will simply be cheaper. But the quality depends on that. Any tool, whether it's the well-known Cursor, or Google AntiGravity, or GitHub Copilot, or a simple extension for the VS Code development environment. The essence is the same everywhere: you have access to some models and to ChatGPT, and to Gemini, and to Claude, and to Grok, and to anyone else. And for each of these models, you have some limits, say, 50,000 tokens, and this one has 150,000 tokens, and so on. The more powerful the model, the better it generates code. The larger pieces of code, the larger pieces of applications it can write for you and do it without errors, but the more expensive it will be for you. That is, the faster it will consume its limits that are included in your subscription. The simpler, the lighter the model you use, the fewer iterations you need to move simply because it will mess up and hallucinate. But the longer you can go, so to speak. It's like, I don't know, a sprint or a marathon. Run slowly, but far, or fast, but short. In fact, the whole difference in choosing between tools depends on this. Simply where the subscription is more profitable now and that's it. There is no difference. No. That is, there is a huge difference between Lavable and PS44. These are products for completely different niches. Here, you simply download an IDE, download a program that is currently more convenient and cheaper for you. Because in terms of interface, in terms of capabilities, whether it's Cursor, AntiGravity, or anything else, they are absolutely identical because they have the same principle underneath. And essentially, all you buy within the choice of these tools is how many limits each of these subscriptions includes. You will not find information anywhere in the public domain that Google AntiGravity provides, say, so many tokens for Gemini, not three, but Cursor so many. This is purely based on personal experience. That is, I tried Cursor, and the limits are consumed faster there, and they immediately offer you: "Come on, pay for the subscription again." Google, at the moment, at least, has been more lenient with this. So, here is a certain Google Integravity, which needs to be installed on your computer, and you pay a subscription there, which provides limits for some neural networks. Okay. What does the work consist of? The work consists of you telling this neural network what to do. It generates code for you, generates folders, projects, formats all this, but you retain full control at the developer level. What is the overall outcome? The work of a developer five years ago. What was the result of their work? In fact, it's a folder, a folder called "project." Essentially, it's a folder on the disk containing text files called code files. All these folders and files within them are generated by the AI, but you, as a developer, still have full access to them. You can go in, you can see what it's writing, you can edit it, you can direct it anywhere, in any direction. You can write in different technologies, write in different programming languages. That is, there is absolute freedom here.
So, what are the limitations of this approach? In fact, there is one limitation - it is, of course, the human operator who will be involved in this. This is not a simple thing. And you can't just jump in without knowledge, without understanding. You will still be bombarded with a lot of terms. You will need to perform some tasks, you will, well, I don't know, set something up, run something as an administrator, so that some dependency files are installed, and so on. That is, the entry ticket in terms of competencies is, conditionally, a junior developer. That is, you understand, you understand what a command line or terminal is, you know how to navigate through these folders, directories, you understand how code works, how functions are written, you know why dependencies are installed, how it all looks, how it works. Ideally, you understand what tests are, you understand what Git is, Docker, at least the basics, in order to simply understand what commands to give. The AI is omnipotent and can do anything, but you still need an operator who has at least basic erudition in this area, in the area of development, and some knowledge base. That is, this is not for people from scratch. But there is plenty of information on the internet now to gather and get this level, I don't know, of a junior developer. Therefore, if you are too lazy to collect this information from the internet, like, what is Docker, what is Git, how it all works, how it all connects together, in order to simply form this picture necessary for such development, then please write in the comments, maybe I will shoot a large, large, detailed course where you can go through all these thorns to the point where you simply open Cursor or AntiGravity and, like a brilliant pianist, start typing your thoughts there. How should you work now, again, at the time of the video release, in such services? First, it's still better to choose the stack, the technology yourself. That is, AI doesn't do it very well. It doesn't know in advance what you want to get in terms of, what is your vision for the current scope of work, but for six months, a year ahead. Therefore, the choice of technologies and tools is still best left to yourself. In terms of choosing models. That is, again, in your subscription, there are limits on some models used. Rating. What are the most powerful, coolest models at the moment, you can see on the LM Arena website. That is, now Opus 4.5 from Anthropic is leading there. The model is expensive, so you won't get far with it, your limits will simply run out. But, please, it writes code better than anyone else. Here's the story. Over time, you will simply learn to balance a little. That is, you understand, the task is quite trivial, simple now, you need to fix something. You switch to a cheaper model, simply to consume fewer tokens and have your subscription last longer. I don't know, you've written a lot of large screens, a lot of logic, a lot of actions. Well, it's better to switch to the most powerful model, taking it from, please, SL Arena, to get exactly what is required. Models are getting smarter. They can hold more and more context in their brains and allow you to create more and more significant parts of functionality at once. That is, work is accelerating and accelerating. And if a year ago working with some models was painful, that is, there were constant glitches, constant errors, you constantly had to update everything, redo it, then now, developing some of your applications or your ideas, I don't feel any brakes at all. That is, it simply generates, it simply works, and with a minimal number of bugs. Also, before starting work, I advise you to have a good understanding of the architecture in your head, i.e., what is our database, what data is stored in this database, what is stored locally, what APIs we use. As for, say, correct prompt writing or the fact that you need to formulate tasks in English, because it understands better, maybe that was relevant a couple of years ago, but now it's already, frankly, a bit much. That is, it has become so powerful, it grasps what you want from it so well, that you don't need to bother, write in Russian, write as you are comfortable. They will understand you and do what you need. The only thing is, perhaps, it's still not worth overdoing it, not worth stuffing such huge chunks, break the elephant into pieces, break the task into micro-tasks, micro-functions, again, understand what is happening at that moment, how it is written. in order to give the right tasks to the AI at the right time. That is, not just a huge canvas at once and do it, it won't do it, but precisely thanks to such iterative action, i.e., in small pieces, step by step, brick by brick, you can build and create powerful, cool services and not worry about technologies.
And what to say in conclusion? I honestly didn't think that the programming profession would last so short. I earned my first money in commercial development, programming, in 2020. In early 2022, I quit to switch to freelancing. And as my tool, I chose the no-code constructor Flutterflow, about which there are tons of videos on the channel. Back then, it seemed really super logical. That is, when you do quite similar applications over and over again, it's strange to write some logic pieces from scratch every time. That is, the constructor format seemed super logical, that here, please, we've made some block, I don't know, a registration block, you've developed it once and then just copy-paste it, insert it, and so on. Now, I don't know, in 2025-2026, it turns out that programmers, about whom they shouted from every corner in 2020, that there was a deficit of a million programmers, a billion programmers deficit, everyone needs programmers, Python courses, who needs them, buy them. Those who shouted from every corner. No-code development, which was very recent. This is simply the vanguard. It was taught in various startup courses, accelerators, etc. Look, if you have an idea, you can somehow cobble something together and get results from it. That all this is simply so unnecessary and that now technical erudition is important. Now experience is super important in terms of understanding what to build, how to build. Again, choose technologies, understand the horizon, I don't know, 2, 5, 10 years, maybe, I don't know, someone is even planning for such timeframes now. Now this is important, this is cool, this is something you won't get from courses and won't get, but the technical side of the issue itself is disappearing at an incredible speed. And what recently programmers were considered some kind of elite that no one would ever replace, and this is intellectual labor. And now they have been replaced first. This is wonderful. Drivers are still alive, no questions asked. But for programmers, of course, their chairs have started to shake. And that's great. That is, I started my company for application development. Such a small bureau. We launch applications from idea to release turnkey, including marketing and promotion. That is, seriously, not just development. I launched it with the idea, with the desire to provide a place for the realization of my ideas. What if a person sees some, I don't know, imperfection in the market, some problem that can be solved with some IT tools, applications, websites, that they can come and for some reasonable money and in reasonable terms get an implementation. And I am glad that, it seems, my profession, which I studied at university, and for which I tried to get a job for a long time, is dying, but it is dying with such good, I don't know, with such good farewell words, because this AI, this development allows even more people to realize their own ideas, to embody, create, bring some thoughts to life, and I am sure that we will all only benefit from this. And if the video was useful to you and you want to understand the technology better, you are interested in how applications are launched, or you are interested in, well, precisely no-code development from the perspective of how to gather all this necessary technical expertise to open Cursor and not close it again, but sit down and do something cool there, then subscribe to the channel and write about it in the comments. I will shoot a video, it will be useful for everyone. Thank you for your attention and see you in new videos. Bye.