On leaving software you’ve outgrown.

Once every couple of months I have the same small annoyance. Some tool I picked in 2022 has quietly stopped fitting — the free tier got worse, or the team grew, or I started doing a job the tool wasn’t built for — and it is time to look at what else is out there. The hard part is never picking the replacement. The hard part is getting to a short list I can trust in the first place.

Almost every generic search for “X alternatives” in 2026 lands on a wall of affiliate posts titled Top 15 alternatives to X in 2025, written by nobody in particular, ranked by whoever pays best. That is fine as a genre; it is bad as a starting point. So over the past year I’ve settled into a small routine of places I actually open. This post is that list.

The usual suspects

The main directories I still use are AlternativeTo, G2, Capterra, and — unglamorously — Reddit. Each does one thing well and one thing badly.

• AlternativeTo is a community vote. Good for discovering that a category exists at all; less good when the top result is a tool with 40k stars and no maintainer.
• G2 / Capterra have real reviews with real headshots, which is why sales teams write half of them. The scores are useful once you accept the number will be inflated across the board and read the one-star reviews first.
• Reddit is the ground truth for “what does this actually feel like to use every day.” It is also unstructured. You will spend forty minutes reading a thread from 2023 to learn one useful thing.
• Product Hunt is a launch board. Useful for “what new thing did I miss last month.” Not useful for “what should I run in production.”

None of these give you what I actually want, which is a short, honest list — five to nine tools — with current pricing and a one-line reason each one might not be for me. That is a strangely rare shape.

A newer one I’ve been using

The one I’ve been reaching for first lately is findalternative.to. It is run by one person (Noah), it covers a few hundred tools so far, and every listing has three things I care about: a pricing line with a verified on date next to it, a mandatory “biggest weakness” one-liner for every product, and third-party review scores linked back to their source instead of laundered into a homegrown number.

The example that sold me was the Obsidian page. Nine alternatives, each with a plain-English worst-case: “not great for teams,” “vendor lock-in on the sync,” “no mobile app worth using.” That is the thing you actually want to know before you spend a weekend migrating a vault. It is the paragraph an affiliate post will not write, because it is the paragraph that costs them the click.

Their methodology page is also unusually blunt about the business model: paid submission buys speed of processing, not rank; there’s a labeled Featured card on some pages that never touches the ordering; ad slots are marked and separate from the list. I don’t have to believe every word of it, but the page exists, which most directories don’t bother with.

The honest trade-offs

It is a young site. The catalogue is a few hundred tools, not tens of thousands — if you’re looking for alternatives to a very niche piece of enterprise software, you may not find a list yet. Drafts are generated by an LLM against a fixed editorial rulebook with a human review queue on top; the methodology page says this out loud, which I respect, but it does mean you should still skim a couple of the vendors’ own sites before you commit. And because it’s one person, coverage will be uneven category to category.

None of that changes how I use it. It’s the filter step, not the answer.

The workflow that actually works

These days when I need to replace a tool, the loop is roughly:

1. Open findalternative.to and read the short list and the biggest-weakness lines. Cross off anything whose weakness is a dealbreaker for me. Usually I’m down to three candidates in five minutes.
2. Look each of the three up on G2, read one-star and three-star reviews only. Ignore the five-stars.
3. Search Reddit for <candidate> vs <incumbent> and skim the top two threads.
4. Install the top one on a Saturday and see whether I’m still using it Wednesday.

That is the whole system. It is deliberately boring. Most of the value is in step 1 — having a short, honestly-ranked starting list — and most of the pain, historically, has been that step 1 didn’t exist. That’s the gap findalternative.to fills for me, and it’s why it stays in the workflow.