30,000 summaries later: what we learned

TL;DR: YouTubeSummary passed 30,000 summaries. More people are visiting every week, the category mix makes sense, and the goal stays simple: keep it free and roughly break even.

30,000 summaries

YouTubeSummary started as a fun project: paste a link, get something readable, move on with your day.

Crossing 30,000 summaries is a satisfying checkpoint. It also gives us enough data to see patterns in what people watch, what they ask the site to summarize, and what it costs to keep things running.

Overview stats

This is the high-level pulse check: overall volume, recent cadence, and the typical size of a summary. The nice surprise is how consistent the “shape” of summaries stays across wildly different videos.

Velocity trends

The direction is clear: more people are using the site over time. Seeing growth across several time windows usually means it is not just a one-day spike, but steady discovery.

Traffic doubled (and so did the bill)

Over the last two months, visitors roughly doubled. That is genuinely fun to watch.

It also means the expensive part of this project (AI calls) gets hit more often, so the bill follows the traffic. Still: more real usage means more weird edge cases, more feedback, and more reasons to keep improving it.

Category mix (top 10)

Education leads the mix, and the long tail stays diverse. That fits the intent: most visitors are trying to compress time on lectures, explainers, business breakdowns, and other “I want the key ideas” videos.

Cost and token stats (sample with cost data)

This is a small (recent) sample, but it is enough to ground the discussion in reality: summaries have a measurable per-run cost, and token usage swings a lot depending on transcript length and structure.

The extreme outliers are always funny: some videos barely have usable transcript content, others are marathon streams that push everything to the limit.

Keeping it break-even

This project is free to use. Right now, Google ads are the only income stream. They do not fully cover the AI costs yet, but it is close enough to be manageable. The goal is not profit; it is simply to keep YouTubeSummary running without bleeding too much.

Thank you

If you have been using YouTubeSummary, sharing it, or sending feedback, thank you. This project exists because people actually show up and use it.

What’s next

We will keep improving summary quality, expanding category prompts, and making the pipeline more resilient as traffic grows. If you have ideas to help this project break even without turning the site into a mess, I am all ears.