For about 15 years, Evernote was one of the few apps I genuinely loved. I started with notes, then shifted to using it heavily for tasks and reminders as well.
One of the things that made this work particularly well was Evernote’s Android widget. I had today’s and tomorrow’s tasks sitting directly on my home screen. Every time I picked up my phone, they were right there, in my face. I could see what needed doing and tick things off without having to remember to open an app.
That worked extremely well for me. Over the years I ended up with around 1,500 notes and more than 3,000 tasks and reminders in Evernote, and most of that I handled through the widget rather than the app itself. From a product perspective that’s the interesting part. The thing I relied on every day wasn’t really the app. It was a strip on my home screen showing today’s and tomorrow’s tasks. And I was perfectly happy paying for it, because it solved my problem exactly the way I wanted it solved.
Then it started to feel stale
Over the years, though, the cracks started to show. Nothing dramatic. Just lots of small things. Adding a task took more clicks than it probably should’ve. Workflows weren’t terrible, but after thousands of repetitions, every extra tap stood out. What really stood out was that nothing seemed to change.
The product felt stale, and that’s hard to miss when you make products for a living. I started glancing at other task managers and note-taking solutions. Not because I was desperate to leave, but because I was curious if anyone had built something better. I even wrote about the widget’s rough edges a couple of years back.
But the real problem wasn’t finding something better. It was leaving. I’ve poured 15 years into Evernote. Thousands of notes and tasks. Exporting was possible, but untangling it all felt massive. As a developer, I could have built a migration tool, but that one-week project always slipped to someday. So I stayed. Not because the product was getting better, but because leaving still felt harder.
Then Bending Spoons arrived
When Bending Spoons acquired Evernote, I was actually hopeful. They talked about investing in the product and for a while it felt like maybe the app I relied on had a future again.
That’s not how it went. They were building outward rather than inward. New features arrived, including the oh-so-important AI additions, while the small frustrations I dealt with every day stayed exactly where they were. We wanted different things from the product.
I’m not saying that their strategy was wrong. It just wasn’t what I had hoped for. If I’d been running a product with a loyal user base, I’d probably start by fixing the rough edges those users hit every day before broadening the product. And when the subscription price increased significantly on top of that, I finally decided it was time to leave.
Getting my data out
I’d already decided where I was going. Todoist for tasks, Notion for notes. My first instinct was API. Simple and straightforward. Evernote had an API, Todoist and Notion had APIs, so as a developer this seemed like the obvious route.
Except Evernote wasn’t issuing new API keys. From Evernote Support, 8 April 2026:
Evernote API Key requests are temporarily suspended while we complete some maintenance on the platform.
I waited about a month to see if that would change. It didn’t. So I moved on to exports.
Evernote lets you export your data into its own data-readable format in ENEX. That’s where AI actually changed everything for me. Instead of building a migration tool from scratch, I worked with Claude to understand the export, separate notes from tasks and reminders, transform the data, and then connect everything to the respective APIs on the other side.
Tasks and reminders went into Todoist. Notes went into Notion. What surprised me wasn’t the code itself. I knew this was technically possible. The difference was how little effort it required from me. A project that I could probably have built myself in about a week became something I finished in roughly an hour.
And after 15 years, I was out.
AI changes the economics of these little projects
What surprised me most wasn’t that AI could do the migration. It’s that it made the migration worth doing. I’ve always been able to write one-off tools like this. Most developers can. The problem is that they’re rarely worth the effort.
Spending a week building something you’re going to run once isn’t particularly appealing. So it ends up on the backlog, and in my case, staying there for years. AI changes that math. It lowers the cost of solving problems that only need solving once.
That’s a very different use case from asking AI to “build my whole app”. It’s taking something I already know is technically possible, but isn’t worth my time, and reducing the cost enough that suddenly it is. It’s the same shift I wrote about in Is the SaaSpocalypse real, or is it just a squeeze?, just applied to my own backlog instead of the software industry.
The setup works. My notes live in Notion, my tasks in Todoist, and I’m getting more done than I have in years. I also stopped paying hundreds of dollars a year for a subscription I wasn’t excited about.
All it took was an hour on a train, and some AI.
Leave a Reply