How Dirr verifies and scores

Where the numbers come from

We read each tool's public, logged-out pricing and feature pages, the same pages you would see. We do not accept terms, log in, or scrape anything personal. A crawler captures the page weekly; a model normalizes the raw facts into a comparable shape.

Every cell we publish stores three things behind it:

Here is that provenance for every tool in the dataset right now, so you can check our work. Nothing here is hand-typed from memory.

The verify-live rule

Some numbers move often, per-credit rates and promotional entry prices most of all. When our confidence in a cell drops below threshold, we do not publish a number we cannot stand behind. We show verify live and link you straight to the vendor's current page.

We would rather show you less than show you wrong. A flagged cell is an honest "check this yourself," not a gap we are hiding. Stable binary facts, model support, self-host, MCP, are far less volatile than exact dollar figures, so we lead with those and treat volatile prices with caution.

How the scoring works

Rankings are not opinions. Each use-case applies a transparent, weighted rubric to the verified facts. The weights are fixed and published, so two people running the same inputs get the same order. Here is the default weighting for a general "best no-code builder" ranking:

Different pages use different weights, a "cheapest" page weights price higher, a "self-host" page weights it as a gate. The weighting for each page is fixed and auditable. We never re-order results by who pays us.

Affiliate disclosure, in full

Dirr is reader-supported. Some links to tools are affiliate links, which means we may earn a commission if you subscribe after clicking one. That is how we fund the weekly verification work.

The order you see is the order the rubric produces from verified data. If a tool that pays us ranks below one that does not, that is the rubric working as intended. Dirr is run by an independent publisher based in Sweden.

Found an error?

One wrong number costs us your trust, so we want to know. If a figure looks off, tell us the tool, the cell, and what you are seeing, and we will re-verify within a day.