The blocklist
What is refused, what is not, and what a refusal costs you.
A blocklist runs on every submission. If your word matches, it is refused, it does not enter the manuscript, and you are told why. Matching runs on a normalized form of your word — case folded, with accents, look-alike letters, and number substitutions resolved — so that a respelling of a listed term is the same term.
You get one refusal a day
The first blocklist match each day is a warning. Your word for that day is still yours and you can try again immediately. A second refusal on the same day spends it. The count resets at 00:00 UTC with everything else, and nothing carries over between days.
The reason the second refusal costs the day is that free refusals turn the blocklist into a search box: someone testing what they can get past it would have unlimited attempts and would eventually find a spelling that works. A budget of two attempts a day is still bounded, and it spares the much larger group of people who typed something careless once.
What you are owed in exchange is warning before you submit rather than after: this page is public, there is a confirmation step on every word, and the first refusal states plainly what a second one costs.
If a refusal spent your day and you believe it was wrong, you can appeal it. A warning costs nothing, so there is nothing to restore — but you can still report a blocklist entry as over-broad through the same form. If the entry was wrong, it is removed or narrowed and your day is returned. If it was right, we say so plainly.
What is on the list
Four categories, and only these four.
- Slurs. Racial, ethnic, religious, sexual-orientation, gender-identity, and disability slurs, in the languages we can source lists for, including character-substitution variants.
- Targeted harassment. Explicit threats, and terms whose function is to attack a person.
- Doxxing vectors. Anything shaped like a phone number, street address, email address, or government identifier.
- Spam. Web addresses, domain names, and terms that exist only to advertise.
What is deliberately not on the list
Profanity. Political terms. Names of public figures. Words that are rude, ugly, depressing, off-topic, or that ruin the sentence someone else was building. Crude words belong to the crowd’s vocabulary, and removing them would make this a curated text rather than a collective one.
The blocklist is for words that could harm a person, not for words we dislike. There is no review of accepted words for quality, taste, sense, or agreement with the surrounding text. The crowd writes what it writes.
How the list is governed
Every entry is public: the term, its category, and the date it was added. Entries added during November are logged with who added them and why, and that log is published with the archive. Adding a term affects future submissions only — words already written stay written. The blocklist never removes anything from the manuscript.
Show the list of blocked terms
The list includes slurs and other terms chosen for their capacity to harm. It is published so the rules are knowable in advance, not to be read for its own sake.
Phase 1 renders the `blocklist` table here — term, category, and date added, ordered by category. It is not searchable and it is not paginated for browsing.
Hiding a word
Separately from the blocklist, a word already in the manuscript can be hidden. That happens in one situation only: a word or sequence of words being used to harm a specific person. The row is never deleted, the position is never reused, and the manuscript renders a visible gap where it was. Every hide is recorded in the removal log.
Any reader can report a word. Reports go to a queue for human review and never hide anything automatically.
If the blocklist refused a word it should not have, or you need to reach a person about a word without signing in, write to support@nevergreengames.com. There is no promised response time — see the removal log for why we would rather say that than name one we would miss.