P.S: you can read this post on my Medium too here https://medium.com/@muhammadarifineffendi/not-sure-what-saas-to-build-try-tinkering-with-the-basics-d176a92faf48!
Lately, I’ve been banging my head looking for SaaS ideas and getting envy with all those successful people making thousands or even hundreds of thousands of dollars with their SaaS.
To me, it seems like those successful people on X has great ideas, great executions, and here I am just a worthless dev couldn’t even build a simple thing.
Have you ever feel the same? Getting this huge impostor syndrome, and fear of not making it?
Or maybe feeling like you’ve been chasing multiple ideas for your SaaS, but always ended in a roadblock.
Well, me too. And in this post, I want to tell you how to get on with it and just START BUILDING WITH THE BASICS.
Exploring the Basics - not basic programming of course
Yep, you read that right.
Last week, I’ve come to realization that instead of keep thinking of the next $10k/month SaaS ideas, and not building anything since I fear things might be failing, I should just try to start from the basic.
By “The Basic”, I’m not saying you should start with basic programming like Array, for-loop, etc. But, the basic of multiple AI technologies that you can find now.
The easiest way is by checking all the available APIs from OpenAI in their API Docs, then start building with one of the API you found there.
For example, when I’m writing this blog post, OpenAI got several Platform APIs for:
- Chat Completions
- Streamings
- Audio
- Embeddings:
- Images
- Realtime
- etc
Pick one thing, then explore the simplest use case you can build with it.
Start simple - focus on backend logics, do minimum frontend works
After you choose whatever API that you want to try, the next part is to focus on backend logic first and do minimum amount of frontend works.
I know that there are multiple platforms like Lovable, Bolt, v0, that can help you build cool looking UIs. But, JUST STOP FOR NOW.
Do not think about the UI, start thinking about the LOGIC FLOW that you want to implement. You can even do UI without any styling, just focus on building the backend functionalities first for now with the API.
If you got too many fancy UIs, you will most definitely feel overwhelmed later. Been there, done that. It’s not a good feeling.
That’s why, just build the backend function first! Make everything works with the simplest UI, then improve the UI later.
An example - building AI transcriber app with OpenAI Whisper
Starting with the simplest STT idea
As an example, last week I decided to explore about building a simplest app with OpenAI Audio API, with Whisper model.
The simplest use case I can think of is building a simple AI Transcriber App with these three core features:
- record audio voice from web app
- transcribe the recorded audio with OpenAI Whisper speech-to-text (STT) model
- show the formatted transcribed audio in the UI
That’s it, no fancy real-time audio, AI agent, and all those complicated things.
With just this idea, I’ve managed to learn so much about:
- audio chunking
- implementing speech-to-text with OpenAI whisper
- implementing transcription with Next.js server action
- and transcription formatting with OpenAI
gpt-4o-mini
Here is a screenshot of the App I’ve built:
More ideas for AI transcription
Then with the basic understanding of how to build with the STT API, I started to get more ideas about what to build.
Several ideas I got, for example:
- Real-time meeting note & summarization
- Tool for repurposing Podcast into written contents for LinkedIn, X, Threads, blog post, etc
- Tool that take unedited video, increase sound quality, add edited list of points automatically when people explaining things in the video, etc
- AI Debate Coach
- Transcription as a Service for Niche Markets
- etc
So, just starting from the basic will give us so much confident that we can build things. Especially when we can make all the backend run as we want and doesn’t get overwhelmed by UIs, it will feels a lot better.
What should you do now?
Just build the simplest stupid app without overthinking things
Well, pick one API from OpenAI or maybe Mistral, Gemini, etc and decide on the simplest thing you can build.
For example, let’s talk about the new Mistral OCR API. There are a lot of things you can build with it, and lots of use cases we can use it. But for simplicity purpose, an idea you can try is to build a simple app where you can:
- upload pdf only pdf file on the app
- process the pdf with mistral API
- then show the markdown result in the UI with the simplest styling possible
Things you will learn along the way
When building that simple PDF tool, you might be learning new things such as:
- how to upload file to object storage such as AWS S3, Cloudflare R2, Supabase object storage, etc
- how to use Next.js server action or API routes to connect Mistral OCR API
- how to use AI SDK from Vercel, or official Mistral SDK, and other libraries
- how to process the returned response from Mistral API, and show it on the UI
- etc
See? Just by focusing on small and simplest stupid use case, you can learn a lot! It will also improve your confidence by just building things, without overthinking about the next $10k/month SaaS.
Just put off the SaaS for now, focus on basic things. Have fun building the basic.
After that, you can start asking ChatGPT, Grok, or other AI to give you some ideas that you can explore from the foundation you’ve build.
Some examples you can try later:
- How do you do PDF chunking and embeddings from the given PDF?
- How do you build a simple RAG app where you can ask anything about the PDF?
- What if you can build quizzes from the uploaded PDF?
- etc
Lots of things you can build with it.
Anyway, that’s it for this post! Hope you got some inspiration, and start building without too much overthinking 😊