ENGINEERING ★ 13 upvotes 2024

Show HN: Launching VideoToPage – How I built it and stack

The Founder's Submission

Direct narrative and submission text authored by @theluk on Hacker News:

Hey there,

so after overcoming imposter syndrom, I decided to launch one of my products to product hunt. https://www.producthunt.com/posts/videotopage

I started thinking about this project right after Vision came out. I thought how cool would it be to have a product demo you recorded turn into a full interlinked documentation with introductions, getting started, code examples and so on.

So I built it and reached a Prototype where this was possible. But it took quite some time to process, and it started to get expensive. Then I did a little bit of SEO research in Ahrefs and found out that there is not even a search demand for that. Then someone told me "give people what they want" and it seams people just look for transcriptions and to repurpose it. So I ditched the screen vision analytics part for now and focused on just the spoken text. Not sure if this was the right decision and if I now removed the main sauce from it. We will see.

The project was built using a stack that allows me to ship within days. It is based on firebase, react and nestjs and nx. I have everything in Typescript. Even though firestore might be tough to work with, I have wrappers around it that help me solve most of the issues. So this is pretty nice.

One thing I am proud of is a think I called AI islands. Imagine a model, like a actual database model, but from a business layer perspective. So in videotopage every model can be extended with AI capabilities. Each model/repository has certain functionality, like editing, deleting, adding, modifying. Now this per-model functionality is exposed obviously to the UI, so that the user can manage his content. Now, all this capabilities are also exposable to an dedicated AI agent. This still does not sound new, but in this case, each model has actually its own AI assistant history, memory and tool access. This way, when you are navigating through the page, you will always see related ai assistance that is focused only on that you see. Also it updates everything you see as if you would do it. This can be sometimes a magical experience, when you look at your list of items with several features in it, and you just start telling the bot what to change and you see it actually change in front of your eyes.

So this AI Island thing is definitely something I will be shipping in all future products.

Retrospective Overview & Key Insights

Distilled analysis of the business dynamics, operational challenges, and strategic takeaways:

In this technical retrospective (2024), @theluk shares practical architectural lessons learned from running, scaling, or migrating production software under lean constraints.

Core Principles for Independent Builders:
1
Defaulting to simpler architectures (such as monolithic frameworks, SQLite, or managed VPS) prevents early engineering bankruptcy.
2
Evaluating true compute and egress costs reveals that high cloud bills can often be slashed by 80%+ through judicious re-platforming.
3
Code you write at 2 AM as a solo engineer must be easy to debug six months later without complex microservice dependencies.
Community Debate Themes:

HN engineers debated the tradeoffs between cloud convenience and raw server economics, database indexing strategies, and resisting premature architectural complexity.

Community Discussions & Critical Reviews

High-signal perspectives, counter-arguments, and peer founder insights debated on Hacker News (1 total discussions):

PE @pedalpete Peer Contributor

Congrats on the launch, looks really great.

The video is well done, but can I suggest giving an example which is more solidly focused on why somebody would turn a video into a webpage?

Your comment above says "give people what they want", but I'm not sure I understand where you found that want. Can you elaborate?

HISTORICAL ERA Post-ZIRP & AI (2023–2026)
COMMUNITY SCORE ★ 13 points
TOTAL DISCUSSIONS 1 comments
TOPICS & TECH STACK: