Why I Built Acloom Notes

Hi, I'm Ashin. My wife is a nursing student.

Nursing classes are long. Three hour blocks are normal, and the professors cover dense material at a fast pace. The kind of pace where you're trying to write notes, listen, and understand at the same time, and you can do maybe two of those three things at once.

My wife would come home after a 3-hour lecture with incomplete notes and spend another two or three hours that night trying to fill in the gaps from recordings and slides. The 3-hour class was effectively a 6-hour class once you counted the post-class processing time. She did this for every class, every day. It was wrecking her.

"Why can't I just pay for an AI tool to do this?" That was the first thought. I'm a software developer. I know AI tools exist. There has to be one for this.

What we tried

We looked at every AI note-taking app on the market.

Most of them were priced for enterprise teams and meetings. Twenty to thirty dollars a month for features she didn't need, optimized for sales calls and board meetings, not for nursing lectures. The student-focused ones were either too basic (just a transcript with no study tools on top) or too expensive for what they actually provided. None of them gave you flashcards, quizzes, and detailed study notes from a single recording.

She tried a few of them anyway. Each one solved part of the problem and left the rest. The transcription was fine, but then she still had to take that transcript somewhere else to generate flashcards. She still had to ask another tool for practice questions. She still had to rewrite the structured notes herself, because the auto-generated ones were three sentences long and missed half the content.

The pattern was the same with every tool. They were great at the one thing they were built for, and bad at the four other things she needed.

The first version

As a developer, my eventual thought was, why not build something ourselves?

The first version was simple. You could upload an audio file, get a transcript, and generate a basic summary. It worked. It saved some time. It was nowhere near enough.

She still had to take the transcript and paste it into ChatGPT along with PowerPoint slides and study resources. She would spend hours explaining what she needed. "Make flashcards from this." "Write practice questions based on these slides." "Create detailed notes that cover everything." Every single time. Every single class.

Then we hit the real problem.

The ChatGPT wall

AI models have limits on how much text they can process. A two hour lecture transcript is huge. ChatGPT would cut off mid-answer, forget context from earlier in the conversation, and she'd have to re-explain everything with new screenshots and re-uploads, even though she had already provided the resources at the start of the session.

She would come downstairs angry. Not at ChatGPT, exactly, but at the fact that the workflow that was supposed to save her time was instead consuming her evening. She had paid for the model. She had pasted in the slides. She had explained what she wanted. And the tool kept losing context and asking her to start over.

This wasn't a fault of ChatGPT. ChatGPT wasn't built for "take this 25,000 word transcript and generate a complete study kit including six different content formats while maintaining context across the whole document." It was built for conversation, and it does conversation well. The problem was the wrong shape for the tool.

What I built next

I started asking her: what would make this actually useful? What's still taking too long? What does the output need to look like for it to replace your manual notes?

The answer was a tool that takes one upload and produces the full study kit. Transcript, summary, structured notes, flashcards, MCQ quiz, short answer practice, long answer practice, essay prompts. All formatted for studying. All generated in one pass. No prompting. No re-explaining.

So I built that. It took a few months of iteration. I'd ship a version, she'd use it for a week of classes, she'd tell me what was wrong, I'd fix it. We did that loop a lot.

The flashcards were the hardest. The first version generated generic flashcards from random sentences in the transcript. The current version picks the concepts that actually matter for an exam, formats them as front and back cards, and groups them by topic. The difference between version one and the current version is whether you'd actually use them or not.

The MCQ quizzes had a similar arc. Early versions wrote bad distractors (the wrong answer choices were too obviously wrong, so the test was trivial). The current version writes distractors that look plausible if you don't really know the material, which is what makes a real practice quiz useful.

She uses it now for every class. Three hour lectures get processed in about 30 minutes. She gets back better notes than she could have written herself in real time, and she has the rest of the evening to actually study.

That's the version that became Acloom Notes.

What I want this to be

A few things I want to be straightforward about.

Acloom Notes is not venture-backed. There's no growth-at-all-costs investor pressure. It's just me building something my wife needed and making it available to other students who need it too. That changes how I think about the product. I don't have to chase enterprise customers. I don't have to add features that look good in a board deck. I can focus on whether the tool actually helps you study.

The pricing is transparent. AI processing has real costs. Every hour of audio goes through transcription, then through note generation, flashcard generation, quiz generation, and so on. Each of those steps has a real cost behind it. The pricing reflects what it actually costs to run the tool plus a margin small enough that students can afford it. I'm not trying to maximize revenue per user. I'm trying to keep the math working while the product gets better.

Support emails come to me. If you email support, you get me. Not a ticketing system. Not an AI bot pretending to be human. I read every message and reply, usually within a day. If something's broken, I fix it. If you have a feature request, I'm the one who decides whether it ships.

The tool stays focused. It's a tool for processing lecture recordings into study material. It's not going to become a calendar app. It's not going to become a chat app. It's not going to become a social network for students. The roadmap is about making the core workflow better, not about adding adjacent features for the sake of growth metrics.

What's next

I'm working on a few things now that I'm excited about.

Better handling of group discussions and seminars (the kind of class where multiple people are talking and the existing speaker handling can get confused). Improvements to how slides and PDFs get integrated into the notes (the current version is good, but there's room to make the cross-references between audio and slide content tighter). More export options so you can take the generated content into whatever study app you actually use.

If you have a feature you wish Acloom Notes did, I want to hear about it. The roadmap is shaped by what real students actually need, and real students email me. The contact form is on the contact page, or you can email support directly.

Try it

If you're a student who records lectures, the 7-day free trial is the easiest way to see if the tool fits your workflow. No credit card required. Upload your first recording and see what comes back.

For a deeper walkthrough of what the tool actually does, Acloom Notes Is for Students Who Record Lectures covers it in detail. If you're more interested in the recording-then-processing study system in general, How to Take Notes from a 3-Hour Lecture Without Burning Out is a good read.

Thanks for reading. If you're a student going through what my wife went through, I hope Acloom Notes helps.

Related

• About Acloom: the full story
• Pricing: plans and the free trial
• Contact: email comes directly to me