Inside the Rule34Dle Character Database: Snapshot, Numbers, and Fandoms
Learn what the Rule34Dle character database contains, where its Archive of Our Own work counts come from, how difficulty tiers are built, and how to use that knowledge in-game.
Every round of Rule34Dle is powered by the same thing: a versioned snapshot of Archive of Our Own character tag statistics. Each entry stores a character tag, its source fandom label, the number of AO3 works using that tag, and an image when one is available.
Looking for the game? Play rule34dle or r34dle to open the homepage and start a run.
Understanding how the database is built will not hand you a winning streak, but it explains why matchups feel the way they do, why some familiar names hide surprisingly low counts, and how to read the numbers with more confidence.
What the character database is
The database is a plain list of character records. For each record, the game stores:
- The AO3 tag name, such as a character's most common name.
- A source label describing the fandom or medium, like "Anime/Manga" or "Video Game" entries.
- A popularity count: the number of AO3 works tagged with that character in the snapshot.
- An image URL, when a public image exists for the character.
The current snapshot contains 3,454 records. Of those, 2,936 characters have images and form the playable pool used in every mode. The remaining 518 records are kept for reference but excluded from rounds, because the game cards would otherwise render without artwork.
A snapshot, not live traffic
The counts are frozen on a published snapshot date rather than queried from Archive of Our Own while you play. This has three practical consequences:
- Scores are comparable. A run played today and a run played next week measure the same data, so leaderboard entries and Daily Challenge results stay fair.
- Numbers can drift from reality. AO3 counts grow every day. A character trending after a new season may already have more works than the snapshot shows.
- Updates are versioned. When a new snapshot is published, the dataset is regenerated and shipped with the game instead of being scraped per visit.
The in-game data page explains this policy in plain language. Rule34Dle is an independent fan project and is not affiliated with Archive of Our Own or the Organization for Transformative Works.
What the counts actually measure
The count is the number of fan works on AO3 tagged with that character. That definition produces a surprising top of the table, because generalized tags outcount any individual character:
| Snapshot entry | Count | Type |
|---|---|---|
| You | 6,161,749 | Reader-insert tag |
| Original Characters | 1,066,592 | General tag |
| Reader | 802,270 | Reader-insert tag |
| Original Female Character | 601,330 | General tag |
| Harry Potter | 562,350 | Highest named character |
Reader-insert and original-character tags sit far above every named character, which tells you something real about fanfiction: a huge share of works are built around the reader or new characters rather than a specific canon cast.
Below the famous names, the middle of the database is dense with characters in the 4,000–5,000 range, such as Pieck Finger from Attack on Titan or Victor Stone from DC Comics. These mid-tier entries are where Hard-mode matchups live, and where fandom knowledge starts to matter more than name recognition.
Fandom coverage
The snapshot carries 665 distinct source labels, mixing named franchises with broad categories like video game, book, and real-person fandoms. The largest franchises by character count include:
| Source | Characters |
|---|---|
| My Hero Academia | 96 |
| Genshin Impact | 86 |
| Harry Potter | 72 |
| DC Comics | 62 |
| Naruto | 60 |
| A Song of Ice and Fire / Game of Thrones | 59 |
| Marvel (Comics/Movie) | 57 |
| Haikyuu!! | 57 |
| One Piece | 46 |
| Attack on Titan | 43 |
Long-running anime and comics universes contribute deep rosters, because years of fan activity accumulate works for even secondary characters. Newer but massive fandoms, like Genshin Impact, get there through sheer volume of active writers.
How difficulty tiers are built
Difficulty in Rule34Dle is derived directly from the count column, using three fixed ranges over the playable pool:
| Tier | Popularity count | Characters |
|---|---|---|
| Easy | 20,000 and above | 295 |
| Medium | 5,000–19,999 | 1,005 |
| Hard | Below 5,000 | 1,636 |
The pools are deliberately uneven. Hard is the largest tier, packed with side characters and smaller fandoms, which is why Hard mode shows three cards instead of two: with less famous characters, an extra comparison gives you one more piece of evidence.
The modes draw from these tiers on a fixed curve. Endless mode uses Easy for rounds 1–5, Medium for rounds 6–10, and Hard from round 11 onward. The Daily Challenge uses Easy for rounds 1–3, Medium for rounds 4–7, and Hard for the final three rounds.
Using the database in gameplay
A few database facts translate directly into better decisions:
- Named characters rarely beat franchise anchors. A protagonist from a decades-old mega-franchise usually outcounts a beloved character from a niche series.
- Longevity compounds. Fandoms that have been active for many years accumulate works for supporting casts, so an unfamiliar name from a classic series can quietly outcount a familiar face from a recent hit.
- Shared universes inflate counts. Comics characters appear across countless crossovers and alternate universes, which pushes their totals up.
- The visible count is calibrated evidence. Once you know the tier boundaries at 20,000 and 5,000, the revealed number tells you which tier you are in and how much surprise is still possible.
When a round reveals its answer, the card shows the character, its source label, and the exact count. Treating those reveals as study material is the fastest way to calibrate your instincts against the snapshot. The seven strategy habits turn these database patterns into a repeatable decision process, and the rules overview covers every mode and scoring formula if you are still learning the game.
Spotting mistakes in the data
Snapshot data can contain errors: a wrong image, a mislabeled source, or a count that looks off. The game does not yet have a dedicated reporting flow, but you can flag problems through the guestbook, and reports are used to correct entries when the dataset is regenerated.
Character database FAQ
How many characters are in the Rule34Dle database?
The current snapshot holds 3,454 records, of which 2,936 characters with images appear in gameplay.
Is the data live?
No. Counts come from a versioned snapshot of Archive of Our Own tag statistics, so results stay comparable across players and over time.
Is Rule34Dle affiliated with Archive of Our Own?
No. Rule34Dle is an independent fan project, not affiliated with AO3, the OTW, or the original AO3dle site.
Can I browse the full database?
There is no character browser yet. The most reliable way to explore the data is to play: every reveal shows the character's source label and exact count.
The database is the quiet backbone of every mode. Play a few rounds, watch the reveals, and let the snapshot teach you how fanfiction popularity really distributes across fandoms.