How we rank anime — the method behind Anime Codex, and how to use it

Most anime rankings are one of two things: an average of user votes (MyAnimeList) or a popularity list (a YouTuber's top 10). Both are useful. Neither answers the question fans actually argue about: what makes a show great, and does this one earn its reputation?

Anime Codex is my attempt at an answer — one consistent critical rubric applied to 190+ widely-known anime, with every score you can sort, filter, compare, and disagree with. This post explains why I built it, exactly how the ranking works, and how to get the most out of every tool on the site.

The problem with how anime gets ranked

A MyAnimeList score is a popularity-weighted average of millions of votes. It tells you how many people liked something and roughly how much — but it has no structure. It can't tell you why a show scores an 8, which part is carrying it, or whether the hype outran the writing. Two very different shows land on the same number for completely different reasons, and you can't see any of it.

A YouTube tier list has the opposite problem: it's one person's opinion, frozen in a video, impossible to sort or interrogate. Great for a hot take, useless as a reference.

What's missing is a method — a single standard, stated openly, applied the same way to every show, that you can actually argue with.

The rubric: six criteria, stated openly

Every show in the catalogue is scored 1–10 on the same six criteria:

• Story & narrative (25%) — structure, pacing, payoff, ambition
• Character writing & growth (25%) — depth, change earned by the story
• Themes & emotional resonance (15%) — what the show is actually about
• World-building & power-system uniqueness (15%) — rules, consistency, originality
• Animation & direction (15%) — craft, key moments, directorial vision
• Cultural impact (5%) — influence and longevity

A 10 means "definitive of the medium" and is used sparingly. A 7 means "good but flawed." Below 5 means real, nameable weaknesses. Crucially, the weights shift by demographic — a shoujo romance isn't judged against a shonen battle series; each show is measured against the best of its own kind. You can read the full weighting in the methodology.

Why AI — and why that's not the gimmick

Here's the part people assume is the trick: yes, the scoring is generated by an AI (Claude). But the AI isn't the opinion — it's the consistency.

I wrote the rubric. I chose the criteria, set the weights, and defined the scale. Then every show runs through the same prompt with the same anchors and the same demand: cite specific arcs, episodes, and directorial choices, not generic praise. The taste is mine; what the machine provides is the patience to apply one critical lens to hundreds of shows without drifting — something no human reviewer realistically does. Every ranking is stamped with the model and prompt version, so the catalogue stays internally consistent and re-runs cleanly whenever the rubric is revised.

Where the Codex disagrees with the crowd

This is the most interesting page on the site: Codex vs MAL lays out the biggest gaps between our rubric and the MyAnimeList consensus, in both directions. Shows the crowd overrates. Shows it sleeps on. The disagreements are the point — they're where a structured method earns its keep, and where the best arguments start.

How to use Anime Codex

The rankings are the foundation; the tools are how you actually use them.

Browse, filter, sort, search

The catalogue is the home base: every show ranked by weighted score, filterable by streaming platform, dub, status, decade, and genre, sortable by overall or any single criterion. Searching is instant.

Compare shows head-to-head

Compare puts two or three shows side by side, criterion by criterion, with an opinionated verdict. Settle the debate with the breakdown, not vibes.

Tune the rubric to your taste

Disagree with the weights? Tune lets you drag the six sliders and watch the catalogue re-sort live. Think animation matters more than we do? Slide it up. It's the fastest way to see how much the weighting drives the result.

Analyze the whole catalogue

The analysis suite reads the data for you: auto-computed Insights (biggest divergences, strongest genres and studios, whether buzz tracks quality), a Tableau-style Explore canvas where you plot any two metrics, plus quadrants, leaderboards, and a per-show radar. No spreadsheet required.

Build and share a list

Make your own ranked or tier list from the catalogue with Lists — add notes, share a link, and every pick links back to its Codex ranking, with its genre, airing status, next episode, and where to watch.

Keep up with what's airing

A free seasonal calendar, weekly airing schedule, upcoming lineup, and anime news — with Codex score badges on the shows we've ranked.

Tools for creators

If you make videos or posts about anime, there's a free creator toolkit: ready-to-make content angles, a "back up your take" data card, embeddable score cards, and a shareable creator hub.

What's free, and what's $7.99

Most of the site is free: a preview of the catalogue, the analysis tools, Codex vs MAL, Tune, the calendar and news, lists, and the creator toolkit. A one-time $7.99 unlock opens the full catalogue, side-by-side comparisons, and the design studio. No subscription, no account — email-based access, buy once.

FAQ

Is this just another anime list? No. The rubric and weights are published, the scoring is applied consistently across every show, and you can re-weight it yourself. Lists tell you what's good; the Codex tells you why.

Is the AI just making up scores? The AI applies a fixed, human-written rubric and must cite specific evidence for each score. It's the consistency engine, not the opinion. Every ranking records its model and prompt version.

Why do some shows score far below their MAL rating? Because the rubric measures critical execution, not popularity. When they diverge, it's usually informative — see Codex vs MAL.

Do I need an account? No. Access is by email magic link after a one-time purchase. No passwords, no subscription.

The whole point

You're going to disagree with some of these rankings. Good — that's the design. The rubric is published, the weights are yours to change, and the disagreements are where the real conversation lives. Browse the catalogue, find the one you think we got most wrong, and argue with the method.