Can you start an AI side business earning $50,000 a month with just a few tools? I've tried out these ten categories of popular tools.

“Can a side hustle in AI making $50,000 a month really be set up just by stacking a few tools?”

As AI tools become increasingly popular, many people are starting to focus on how to use these tools to achieve rapid commercial monetization. Renowned full - stack developer and heavy AI tool user Ras Mic conducted an in - depth analysis of ten popular types of AI tools on the market in the latest podcast. From n8n, Lindy, Claude Code, Devin, Code Rabbit, to Bolt, Lovable, VAPI, MCP, and the application of Vibe Coding tools, he detailed the real uses, suitable user groups, achievable effects, as well as hidden thresholds and misunderstandings of these tools.

Which are the high - threshold tools that “programmers can use proficiently, but non - technical people had better avoid”? And which are the new paradigms that can truly enable “newcomers to directly achieve commercialization”?

As various AI tools continue to lower the technical threshold and shorten the product development cycle, who can turn ideas into reality faster? Is it the technical CTO who “questions first and then acts”, or the product - oriented CEO who “tries first”? Therefore, on the basis of evaluating these tools, Ras Mic also reviewed the topic of “making $50,000 a month from an AI side hustle”, analyzing the challenges and opportunities within.

More important than the “AI tool boom” is that these tools are bringing a brand - new way of innovation and thinking. If you are a developer, a product person, or interested in starting a business with AI tools, you might as well take the time to read this compilation (since most of the time it is Ras Mic's wonderful narration, this article is presented in a non - dialogue form with slight abridgments).

1  n8n: Not very useful for developers

If you don't use modern tools much, for example, if you're not a heavy user of Google Drive, don't use Slack often, haven't done any automation, and haven't been exposed to Zapier, this tool is basically useless to you.

Moreover, I've seen it going viral on Twitter. It's almost become the new generation of “e - commerce startup” scam. I really... If n8n has an affiliate program, it would be a repeat of Shopify. Now everyone is talking about automation. In this sense, I think it's a bit overrated.

From a developer's perspective, you'd think, “I can write these things myself. I don't need your system.” Or even if you're too lazy to write them yourself, you can just ask ChatGPT for a JSON and copy - paste it.

But for most people, when they copy the automation processes shared by others, they can't get them to run at all, or when something goes wrong, they don't know what to do. So for most people, this tool is overrated, but I think it's underrated for semi - technical or technical users. By “technical organizations” I don't necessarily mean those who write code, but those who understand some technical principles.

2  Lindy.ai: Great at marketing

The Lindy.ai tool is also quite interesting, but it's more marketing - oriented.

Actually, n8n also has a lot of marketing elements. Lindy's advantage lies in its many templates. I think it's underrated.

It has about hundreds of templates. For example, if you want to set up an automatic email outreach agent, you can just click a button to copy and paste the template and use it directly. I think the great thing about this is that the templates can inspire ideas. There are some use cases I didn't think of before, but after seeing the templates, I can apply them.

3  Claude Code: High threshold, but worth a try

Claude Code is really seriously underrated.

Although it's already popular all over the Internet, I really think it's extremely powerful. If I could only choose one AI tool and AI didn't progress in the future, I'd definitely choose Claude Code.

There are many online videos teaching you how to use prompts to create Markdown files, write PRDs, and feed context... These skills are of course important and make people think more like product managers, which is great. But Claude Code shows a completely different possibility: it can build a “task agent” that can track tasks, plan processes, execute operations, and write tests by itself... Most programmers or independent developers don't like writing tests, and I do too, but tests are really important. Claude Code can automatically write tests for you.

If you're a developer and haven't tried Claude Code yet, you're really missing out. Even if you're not a developer, there are many teaching videos online. Just find one, install a development environment like Cursor or Windsurf, and pair it with Claude Code, and you can experience “intelligent coding” in the true sense.

In my opinion, Claude Code is the purest and closest tool to the concept of “Agentic Coding” that we've ever seen.

Of course, for non - technical users, even starting Cursor can be daunting.

I know it's difficult. If it were easy, everyone would be using it. Although I'm saying this from the perspective of “these tools are easy for me”, if you've never tried, you'll really miss a lot. You should at least try once. If you get stuck, it doesn't matter - you can use another AI tool to solve the following problems.

Although these tools weren't originally designed for non - technical people, I still recommend giving them a try.

Another exciting thing is that Claude Code recently released an SDK (Software Development Kit), so many AI platforms can directly integrate Claude Code's Agent without having to develop their own from scratch.

The Claude team has straightforwardly said that they now recommend many developers building AI coding assistants to simply use Claude's Agent. Their exact words were: “We'll easily outperform you.” (Of course, they said it politely)

Seriously, Claude has completely won and established itself in the field of AI coding. I wouldn't be surprised at all if in the future, all “Vibe programming” tools, such as Devin, Code Rabbit, Lovable, Tempo, Bolt, Figma Make, Hostinger, Airtable, etc., all integrate Claude Code's Agent.

After all, these researchers are now worth hundreds of millions of dollars. Who wants to waste money “reinventing the wheel”?

So I think, although Claude Code has a relatively high threshold for non - technical users, it's still worth a try.

Once you really understand how to use it, the feeling is completely different. I'd place Claude Code in this quadrant: it's not particularly easy to use, and the user experience isn't particularly friendly, after all, it requires using the terminal, which is quite intimidating for many people. But to be honest, no other tool is as powerful as it is.

So I place Claude Code in this position in the diagram. Both technical and non - technical users have seriously underrated this tool.

4  Devin, Code Rabbit: Suitable for building projects from scratch

Devin and Code Rabbit are the two AI tools that are currently closest to being “real assistants”.

The positioning of these two tools is to help you build projects from scratch.

Devin's interface may look similar to other Vibe coding tools, but it's actually designed specifically to integrate with your existing codebase. It has three really amazing features.

First: It has a feature called DeepWiki. You just need to connect your GitHub repository, and it can read the entire codebase and automatically generate detailed project documentation. This is of great significance for people working on formal projects. You don't have to be from a technical background, but after reading this documentation, you can understand the system's architectural principles.

For example, this small project is for creating Apple Wallet Pass cards for community members. Devin automatically disassembled the entire process for me, such as how to generate the Pass cards, how to register the devices, and how to update them. It can also identify that I'm calling the Apple Pass external service and prompt me to click to view the interface documentation and data flow diagram.

It will explain the API call process, such as which endpoint is used for device registration, and clearly show the data formats in and out.

If you're a small team of three or four people working on a serious project, this single feature alone is worth the price of Devin.

The second highlight is that you can start a “Session” to let Devin perform a certain task. You can ask it to fix bugs, add new features, and it can run multiple tasks in parallel. The best part is that it supports deep integration with Slack and Linear.

You can write a product requirement document in Linear and then directly ask Devin to implement it. It will read these requirements, automatically start the development process, and finally automatically generate a PR (Pull Request) and submit it to GitHub.

I once had a very small bug, just a one - line code problem, but I spent hours trying to find it. Later, I used Devin to handle it. It directly wrote the PR for me, explaining which part was modified, why it was modified, and which files were affected. The deployment status was also clear at a glance. I just had to review it and merge it if it was okay.

For me, tools like Devin and Code Rabbit are “practically usable” AI assistant - type tools. If you're a product manager who can't write code but can write a clear functional document, you can definitely use these tools to implement the functions. In my current work, I actually spend less time writing code. I mostly review the results submitted by Claude Code and Devin: judge whether what they've done is reasonable. If it is, I merge it; if not, I send it back for re - work.

The price is also very friendly. Devin's billing model is to pay a one - time fee of $20 to unlock, and then be billed according to usage. It's really worth a try. Although not everyone is suitable for it, you won't know until you try.

Code Rabbit is a similar tool, belonging to the “assistant - type AI Agent”. I haven't fully mastered it yet, but it's currently very good at PR review.

Code Rabbit will automatically review the code when you submit a PR. Developer Code with Antonio once tweeted: “In the past 90 days, I've released 3 courses. Code Rabbit helped me review 77 PRs, left 42 comments, pointed out 66 issues, and provided 61 suggestions. Seriously, can I still be considered 'developing alone'? Install it quickly.”

Another great thing about Code Rabbit is that if it finds a problem, it will compile a prompt for you, which you can directly feed to the AI agent you're using, including the context of the problem, code snippets, etc. Even non - technical users can understand it.

Antonio recently made a clone of Lovable by himself, line by line. You know, Lovable is a product worth billions. And he can replicate it alone. So, if even Antonio says this tool has changed his development process, and he's writing production - level code and can find 66 issues, it's definitely not a trivial matter.

Also, Vibe programming is really cool, and you can quickly create your own products. But if you plan to have real users, you must consider security, payment security, and proper protection of user data. These can't be taken lightly.

In short, I think Devin and Code Rabbit are among the most underrated AI tools at present.

5  Bolt, Lovable, Tempo: Can't replace engineers

For me, Bolt and Lovable are actually very similar. If you haven't used Bolt and Lovable, they're underrated tools; but if you've used them but don't know how to write good prompts, they'll seem overrated; on the contrary, if you know how to provide good prompts and context, they're still underrated.

Once you've used Bolt and Lovable seriously, you'll start to realize that writing software is so difficult. After using them to a certain extent, you'll really feel that the challenges have escalated.

So I think the reason they're underrated is that the initial excitement of getting feedback from writing prompts, that feeling of “opening a new world”, is really exciting. I've seen many people's eyes light up when they experience it for the first time. But when you want to use them more deeply and thoroughly, you'll start to realize that these AI tools aren't omnipotent.

The next big startup wave may not be AI applications, but infrastructure - type companies. For example, Supabase is doing very well. Companies like Supabase and Convex, which offer backend - as - a - service, can package all the underlying details and seamlessly connect these vibe coding tools. I think this is the big opportunity.

Many people say “AI will replace engineers”, but I think these basic services still need engineers to build because it's really difficult. Even though Claude Code is very powerful, there are still some things it can't do.

But for now, Bolt and Lovable have brought great unlocking capabilities and inspiration. I think their “popularity” is reasonable. People who like them definitely have their reasons. And if you haven't used them before, experiencing them now is really like entering a whole new world.

As for which is better, Bolt or Lovable? I think there's no standard answer to this question. It's like Cursor and WindSurf. I'm often asked which is better. They're updated very frequently. In fact, you just need to choose one and stick with it. Of course, you can also use both.

The superiority and inferiority between tools are also fluctuating - Bolt may be better this week, Lovable may be stronger next week, and another tool may take the lead the week after. Unless you're particularly concerned about some exclusive features, these differences aren't that significant.

In short, non - technical users are already living in an era of “tool freedom”.

6  Figma Make: Just riding on the wave

I haven't used Figma Make, Hostinger, or Airtable, and probably won't.

It makes sense for Figma to launch an AI tool. Figma is a core tool for designers, but to be honest, I think it's overrated. Their foray into AI programming tools is somewhat for show. It's not that they really want to compete head - on with Tempo, Bolt, and Lovable. It's more of a market stance. We guess they're trying to ride on the wave of Bolt, Lovable, and Tempo to make the market more interested in their IPO.

Hostinger and Airtable have also each launched AI tools, but I haven't used them. I know Hostinger