Skip to content
Coverage

Coverage

How much of anime history this dataset actually manages, by release year and release season. Every number below is counted from the committed data/, so it describes the dataset as it stands — not what the upstream sources contain.

What is counted

A work is one node that maps to a real anime release: a Season, a Movie or a Special. A Series or Franchise is our grouping and is never counted as a work, so a five-season show contributes five works, not six.

Seasons carry a releaseYear and a releaseSeason; movies and specials carry only a year, so they share one column.

Works by year and season

YearWinterSpringSummerFallFilms & specialsTotalEpisodes
20061124
20111113
20121112
20131175
20141112
20151151
20161125
20171113182
201812348
20192125139
20201321797
20213111676
20221222792
2023344314194
2024442414209
2025333110110
202680691491870
Total9590172032253229

Totals

Count
Franchises1
Series152
Seasons222
Movies3
Works (seasons + movies + specials)225
Episodes3229

Cast

The R2 layer — characters and the staff who voice them — is far thinner than the structure, because it is bounded by what Wikidata (the only CC0 name source the build may redistribute) actually records.

Count
Series with at least one character47 of 152
Characters647
…of those, with a voice actor607
Staff (voice actors)416
Voice-actor links651, all ja

Coverage is uneven in a way that tracks a title’s popularity on Wikidata rather than anything about the show: Slime carries 148 characters and Frieren 43, while most series carry none at all. A series with no cast is normal, not a gap in the build.

Only Japanese voice actors are recorded. Wikidata’s voice actor (P725) statements cover dub casts too, but a language is only trustworthy when the statement says so — so a link is taken as Japanese when the statement is either qualified language of work or name = Japanese, or carries no language qualifier and the actor is a Japanese citizen. Anything qualified to another language is left out rather than guessed at. Other languages can be added the same way, into their own staff/ file.

How to read the shape of it

Winter and Spring 2026 are complete for TV. Every TV anime the pinned sources place in those two seasons is in the dataset — 149 works. Nothing after Spring 2026 is included: an announced but unaired season is left out until it airs, so the dataset never carries a season with no episodes.

Everything before 2026 is there because it is a prior season of something in 2026 (plus the hand-authored Demon Slayer and Fate entries). That is why the earlier years are thin and uneven: 2023 has 14 works not because 2023 was sampled, but because fourteen shows running in 2026 started or continued then. A show like Frieren pulls its 2023 season in with it.

Films and specials are barely covered. Only three exist, all authored by hand. The bulk import is TV-only: of the 580 Winter+Spring 2026 entries in the offline database, only 246 carry an AniList id at all — and the OVA/ONA/SPECIAL long tail, which the builder resolves facts by AniList id alone, is mostly recaps and shorts without one.

Regenerating this table

The numbers come from data/ and nothing else, so they can be recomputed at any time:

import glob, yaml
from collections import Counter, defaultdict

works, episodes = defaultdict(Counter), defaultdict(int)
for path in sorted(glob.glob("data/series/*.yaml")):
    record = yaml.safe_load(open(path, encoding="utf-8"))
    series = record["franchise"]["series"] if "franchise" in record else [record["series"]]
    for one in series:
        for kind, bucket in (("seasons", None), ("movies", "FILM"), ("specials", "FILM")):
            for node in one.get(kind) or []:
                works[node.get("releaseYear")][bucket or node.get("releaseSeason") or "?"] += 1
                episodes[node.get("releaseYear")] += len(node.get("episodes") or [])

for year in sorted(y for y in works if y):
    print(year, dict(works[year]), episodes[year])

GetHealth on the API reports the top-line totals (franchises, series, seasons, episodes) for a running server.