ВЕРСИИ (13)

# NeoModeration 1.7.1 A cloud trial that ends now says so, servers stop spending cloud checks on a refused key, and every site gets the same readable jar. ## Fixes - When the free cloud trial ends, the console and every `/nmod` command that shows the cloud say so, with the way on: create a free account, then a key, and run `/nmod setup <key>`. 1.7.0 called an ended trial a rejected key. A trial the platform stopped says that too. - When a running trial uses up its credits, they say that too. `/nmod trial status` in game gives the button that keeps the trial in a free account, where credits can be added; 1.7.0 sent admins to billing, which needs an account. Whoever opens that link first owns the trial, so it is only ever a button, never console text. - An admin hears once, on joining, when the trial has two days left, has ended or was stopped, or has used its credits, with what the cloud caught and the button that fits. A trial kept in an account shows as kept within about an hour, and one failed answer from the trial service no longer makes the plugin forget what it knew. - After the cloud refuses a key or runs out of credits, chat and map art stop asking for ten minutes instead of asking for every message. `/nmod test` still asks, so a fixed key shows at once. Changing a setting no longer lifts the pause unless it is the key, the endpoint or a timeout; `/nmod reload` always does. - 1.7.0's notes promised far fewer false alarms from everyday chat. That holds for English chat. On 902 clean test lines from players who write in another language, a server whose locale is English catches 223, against 154 in 1.6.6, and 25 with `moderation.offline.trustPlayerLanguage` on. `/nmod doctor` now says when many of the players online play in another language. ## New - `/nmod discord <webhook URL>` mentions once that Moderator Bot, also by Neomechanical, can moderate that Discord server. - `/nmod status` and `/nmod doctor` have a Cloud plan line: the trial's end and the days left, or what to do once it ended or ran out of credits. When that line explains a refusal, the last-check line only says the check was refused. - bStats gets five aggregate charts: chat detections per thousand messages, chat messages a day, how many words are allowed, how many players online play in another language, and whether their game languages are trusted. No message text or player name is sent. - Every site gets the same jar, the one the release checks run. Earlier releases were obfuscated, with a separate build for Modrinth. ## Checks Before release, every check ran on Paper 1.18.2, 1.19.4, 1.20.6, 1.21.11, 26.1.2, 26.2 and 26.3, on Folia 1.21.11, on Spigot 1.21.11 and on Purpur 1.21.11, with a game client chatting on each. After installing, run `/nmod doctor` and `/nmod test badword`. `/nmod mode enforce` makes local rules block and `/nmod cloudmode enforce` does the same for cloud decisions; `monitor` makes either one only alert staff.

# NeoModeration 1.7.0 Server ads blocked on a new install, far fewer false alarms from everyday chat, every message in the player's own language, and staff tools you can run from chat. Local rules still need no account. ## New - Blocks ads for other Minecraft servers on a new install: public IP addresses, any address with a port, `play.` and `mc.` style addresses, free hosts such as `aternos.me`, and "join myserver.net", including "play dot server dot net". Ordinary links, Discord invites, browser games ("lets play skribbl.io") and version numbers stay clean. Let your own address through with `/nmod allow url add <address>`; the first admin to join gets a button for it. - Tells a player why their message, sign, book, item name or command was stopped (`moderation.tellPlayers`). - Every message is written out in English, Spanish, French, German, Portuguese and Turkish, and each player reads them in their own game language (`perPlayerLanguage`). Dates read the way each language writes them, and so do kick and ban reasons you have not set yourself. A prefix you change in your server's language shows for every player. - Starter word lists for Spanish, French, German, Portuguese and Turkish follow `locale` on a new install; choose others with `moderation.offline.languagePacks`. The English list adds 8 slurs with their plurals, and common spellings such as "fvck". - Some words are everyday in one language and banned in another: French "retard" (late), the Spanish name Kike, Swedish "slut" (end), "puto" in Portugal, Danish and Norwegian "fag" (a school subject). They pass on servers in that language. With `moderation.offline.trustPlayerLanguage: true` they also pass for a player whose game language it is, in a message written in that language. A common English word that aims an insult, such as "you", "dirty" or "shut up", still gets the message blocked. The setting is off by default, because every player picks their own game language. - Write a banned word with a space, such as "ass hole", to catch it however it is spaced. - `/nmod cases` pages through older cases, and each case has buttons to resolve or dismiss it, allow the word or mute the player for 10 minutes. - `/nmod mute <player> [duration]` and `/nmod unmute <player>`. A mute covers private messages too (`/msg`, `/tell`, `/r` and any alias the server has for them, such as EssentialsX's `/m`), and `/me`, sending `/mail`, `/teammsg` and mcMMO party chat. Another plugin's command with a shared name, such as Towny's `/t`, still works. - `/nmod discord <webhook URL>` sends alerts to a Discord channel, batched so a raid cannot flood it and with mentions turned off. `/nmod discord test` checks it, and `/nmod doctor` says when Discord refuses alerts and why. - Moderators with the new `neomoderation.moderate` permission can use cases, mutes and tests while settings stay with admins. - During a raid, staff alerts in chat pause. Alerts about one player pause after three in 30 seconds, and all alerts after a dozen in 10 seconds or 30 in a minute. A pause ends after a quiet minute or five minutes, so a raid in waves is one pause. Then one line says how many staff missed, and the console and case history still get every alert. - `/nmod trial` says what a trial sends before anything is sent, `/nmod trial start` starts it, and `/nmod trial status` counts what the cloud caught over the whole trial. - `/nmod surface` is a clickable menu with a button for each setting. - Map art: a map drawn a moment ago is checked at once, an edited map is checked again, and a flagged map hanging in an item frame is taken down without blaming the player who found it; the cloud request names only the frame. Each player can have 30 new pictures checked an hour (`moderation.mapArt.checksPerPlayerPerHour`), so nobody can use up your credits by drawing maps, and the console says when a player reaches it. - Admins hear about a newer release once each time the server starts, when they join. ## Fixes - Fewer false alarms. Everyday words that together spell a banned word pass: "put a torch here", "put a 5 tick delay", and on a Spanish server "vi a dos creepers" or "vino Mari con su novio". So do "let's hit him", "who're you", "took 5 hits" and "455 diamonds". "scam" is no longer a banned word, so players can warn each other about scams. Disguised words such as "5h1t", "f uck", "f'uck", "ni gger" and "k y s" are still caught, and so is a word split into pieces from different languages. - Messages in capitals go out in lower case instead of being blocked, the way the player's language writes them, and only in enforce mode; colours another plugin added stay. Spam is only blocked: it earns no strike and runs no action. - In cloud monitor mode, the trial's default, chat no longer waits for the cloud's answer. A burst of chat waits its turn in a short queue, the cloud is paused only when it is slow or unreachable itself, and a flagged answer that comes too late still alerts staff. `/nmod status` and `/nmod doctor` count the messages the cloud had no time to check because chat was busy, instead of reporting the cloud as down. - `/nmod on` and `/nmod off` replied with a raw message key instead of a sentence. - The update check had told no install about a new version since the repository moved. - A config without `cloudMode` follows `mode` again, and a config without a strikes section no longer turns on a kick ladder. - The same message counts as spam only when it is repeated within 30 seconds, so "gg" after each game is fine. - `/msg <player>` checks only what follows the recipient's name. - Locale files you never edited pick up new and changed messages on update; texts you edited are kept, and a locale file with a typo is left untouched instead of being replaced, with the line to fix named in the console. - Keys and webhook URLs typed into `/nmod setup`, `/nmod key set` and `/nmod discord` no longer appear in the server log. - Chat that another plugin sends for a player, such as `/sudo` or a chat game, is checked by the local rules without holding the server for the cloud. Spigot used to let it through unchecked, and with a cloud key Paper held the server until the cloud answered. - Saving mutes no longer holds the server, and two admins changing settings at once on Folia can no longer damage `config.yml`. ## Upgrading Your config, edited word lists and edited messages are kept. An upgraded server keeps ad blocking off (`moderation.offline.blockServerAds: false`) and, on a Spanish, French, German, Portuguese or Turkish server, its language's word list off (`moderation.offline.languagePacks: []`); the console says how to turn each on. A server whose own language file holds edited messages keeps showing them to every player (`perPlayerLanguage: false`). An unedited 1.5 to 1.6.6 starter list is replaced with the new one, which blocks 8 more slurs and their plurals. Some are everyday words in other languages, such as the Spanish name Kike and Danish "fag" (a school subject). The console names them, and `moderation.offline.trustPlayerLanguage: true` lets players use them when they write in their own language. Players are now told why a message was stopped; set `moderation.tellPlayers: false` to keep 1.6's silence. After installing, run `/nmod doctor` and `/nmod test badword`. `/nmod mode enforce` makes local rules block and `/nmod cloudmode enforce` does the same for cloud decisions; `monitor` makes either one only alert staff. Before release, every check ran on Paper 1.18.2, 1.19.4, 1.20.6, 1.21.11, 26.1.2, 26.2 and 26.3, on Folia 1.21.11 and on Spigot 1.21.11, with a game client chatting on each.

# NeoModeration 1.6.6 Configure extra surfaces in game, run in six languages, and click presets, surfaces and cases directly in chat. Local word/link and anti-spam checks work without an account; optional cloud evaluation starts with `/nmod trial` and runs in monitor mode. - Add `/nmod surface <sign|book|anvil|command|all> <off|monitor|censor|block>` (alias `/nmod surfaces`) with tab completion and clickable rows so server owners can arm signs, books, anvil renames and private-message commands in game without editing `config.yml`. - Ship bundled locales for French (`fr_FR`), German (`de_DE`), Brazilian Portuguese (`pt_BR`) and Turkish (`tr_TR`) alongside English (`en_US`) and Spanish (`es_ES`), and resolve short, lowercase or hyphenated locale codes automatically. - Make `/nmod trial`, `/nmod preset` and `/nmod surface` execute on click in `/nmod`, and make preset rows and `/nmod cases` rows clickable for in-game operators. - Record flagged map-art detections in `/nmod cases` and send staff alerts in both monitor and enforce modes. - Align in-game brand colours with `#9ca8ff`, remove retired join glyphs, and migrate unmodified legacy locale strings automatically on startup. - Centralise sibling API endpoint resolution and JSON parsing across usage, trial and update checks. After installing, run `/nmod doctor` and `/nmod test badword`. Review your starter word list. `/nmod mode enforce` controls local rules; `/nmod cloudmode enforce` controls cloud decisions separately. Automated verification covers Paper 1.18.2, 1.19.4, 1.20.6 and 1.21.11.

# NeoModeration 1.6.5 Cloud verdicts and player input handling are more reliable. Local word/link and anti-spam checks work without an account; optional cloud evaluation starts with /nmod trial and runs in monitor mode. - Parse complete JSON scores, including scientific notation. Respect disabled categories and custom thresholds; malformed or missing verdicts count as unavailable. - Cloud outages and busy workers follow failOpen without issuing strikes, mutes or other player punishments. Cloud monitor mode continues to allow chat. /nmod test shows the same outcome. - Remove quadratic local matching for long whitespace and repeated separated-letter inputs. - Scan namespaced forms of configured messaging commands. Block flagged extra-surface content when censoring cannot remove it, including split sign/book content. - Cache map verdicts by saved image content and cloud settings, so changed artwork is checked again. Limit concurrent map scans and cloud chat work. - Protect trial activation from duplicate requests and configuration races; apply its key on the global scheduler and start cloud checks in monitor mode. Reject malformed trial status responses. - Disclose aggregate bStats charts in the public README. After installing, run /nmod doctor and /nmod test badword. Review your starter word list. /nmod mode enforce controls local rules; /nmod cloudmode enforce controls cloud decisions separately. Optional sign/book/anvil/command surfaces remain configurable. Automated verification covers Paper 1.18.2, 1.19.4, 1.20.6 and 1.21.11. Current calendar releases need a separate startup check; this release makes no new automated 26.x chat coverage claim.

# NeoModeration 1.6.4 — Trial Diagnostics, Cloud Category Reporting, and URL Evasion Coverage This release surfaces `/nmod trial` and `/nmod trial status` across all in-game help, startup logs, and diagnostics, reports the specific cloud category that triggered a verdict (`platform:<category>`), and closes dot-obfuscated URL evasion (`discord[.]gg`, `discord dot gg`) in default URL matching. ## New and Improved - **In-Game Trial Diagnostics (`/nmod trial status`):** Check active 14-day evaluation trials, remaining days, expiration timestamps, and workspace claim links directly from Minecraft. - **Surfaced Trial Onboarding:** `/nmod help`, `/nmod doctor`, `/nmod status`, `/nmod test`, and startup logs now guide operators to `/nmod trial` when no API key is configured, and existing locale files automatically migrate unchanged legacy strings. - **Specific Cloud Category Attribution:** Staff alerts, `/nmod test`, and `/nmod cases` now report `platform:<category>` (for example `platform:harassment` or `platform:sexual`) instead of a generic `platform` reason when the cloud flags chat. - **Dot-Obfuscated Link Detection:** Default banned URL matching now normalizes bracketed and word-substituted dots (`[.]`, `(dot)`, ` dot `) so disguised invite links are caught without requiring wildcard rules. ## Verify and upgrade Run `/nmod test badword` to preview the bundled local rule without executing actions. Run `/nmod status` and `/nmod doctor` to inspect configured checks and cloud health. Local matches block on a new install; `/nmod mode monitor` previews them through alerts, while `/nmod mode enforce` restores blocking. Cloud decisions start in monitor mode and only block after `/nmod cloudmode enforce`. Run `/nmod trial` to begin a 14-day evaluation trial of context-aware cloud protection, or `/nmod trial status` to inspect an active trial. **Existing servers keep their saved `config.yml`**, including action lists, strike settings, modes, category toggles, word rules and API keys. Review the starter English list for your server. This release has no measured effect yet on installations, repeat use, cloud activation, sign-ups or revenue.

# NeoModeration 1.6.3 — Minecraft Threat Intelligence and Instant Evaluation Trial This release introduces direct server-side threat detection for obfuscated IP addresses and advertising patterns, along with a zero-friction evaluation bridge for Neomechanical cloud moderation. Server administrators can now activate a 14-day trial of cloud scanning directly with `/nmod trial`. ## New and Improved - **Instant Evaluation Trial (`/nmod trial`):** Server owners can test context-aware cloud moderation immediately without leaving Minecraft or manually generating credentials upfront. Evaluation activations are authenticated via signed plugin handshakes with server-side expiry. - **Obfuscated Advertising and IP Filtering:** Enhanced detection for evasive server IP formats, disguised dots, spaced digits, and port evasion in chat and signs. - **Trial Status and Expiration Reporting:** `/nmod trial status` and `/nmod doctor` display remaining trial days, credit quotas, and direct links to claim or extend workspace keys. - **Hardened Handshake:** All evaluation requests use tamper-resistant install identities and timestamped HMAC verification to protect trial access. ## Verify and upgrade Run `/nmod test badword` to preview the bundled local rule without executing actions. Run `/nmod status` and `/nmod doctor` to inspect configured checks and cloud health. Local matches block on a new install; `/nmod mode monitor` previews them through alerts, while `/nmod mode enforce` restores blocking. Cloud decisions start in monitor mode and only block after `/nmod cloudmode enforce`. Run `/nmod trial` to begin a 14-day evaluation trial of context-aware cloud protection. **Existing servers keep their saved `config.yml`**, including action lists, strike settings, modes, category toggles, word rules and API keys. Review the starter English list for your server. This release has no measured effect yet on installations, repeat use, cloud activation, sign-ups or revenue.

# NeoModeration 1.6.2 — First-run status and cloud coverage corrections This release corrects what a new server owner sees after installing 1.6.1. The intentional block-only policy no longer appears broken in `/nmod doctor`. Startup, status, doctor and privacy now use the same configured-coverage check, so they do not describe disabled checks as active. These are reporting corrections; they do not change an existing server's saved policy. ## Fixed - `/nmod doctor` treats an empty extra-action list as healthy: local rules can still block a matched message without muting or clearing chat. It names local and cloud modes separately. - `/nmod status`, startup output and `/nmod privacy` describe only checks that are configured to run. bStats coverage charts use the same calculation; cloud account health remains a separate question from configured coverage. - With all cloud text categories disabled, chat no longer sends a cloud text request. Key-gated map scanning is independent and unchanged. - `/nmod test <message>` respects disabled chat scanning and correctly previews local censoring. The command now explains that rate, caps, repetition, command and map checks need live context; a one-message cloud preview cannot demonstrate conversation context. ## Verify and upgrade Run `/nmod test badword` to preview the bundled local rule without executing actions. Run `/nmod status` and `/nmod doctor` to inspect configured checks and cloud health. Local matches block on a new install; `/nmod mode monitor` previews them through alerts, while `/nmod mode enforce` restores blocking. Cloud decisions start in monitor mode and only block after `/nmod cloudmode enforce`. **Existing servers keep their saved `config.yml`**, including action lists, strike settings, modes, category toggles, word rules and API keys. Review the starter English list for your server. This release has no measured effect yet on installations, repeat use, cloud activation, sign-ups or revenue.

# NeoModeration **Chat moderation that works on a fresh install.** A starter English word list, configured URL rules, and anti-spam run locally with no account required. Fresh installs enforce local detections immediately. Review the list and actions for your server; add optional cloud moderation for context-aware chat checks and NSFW map-art scanning. **Test before changing your server:** `/nmod test <message>` previews a decision without taking action. Run `/nmod mode monitor` for alert-only local evaluation. Cloud decisions begin in monitor mode and need a separate `/nmod cloudmode enforce` before they can block. ## What it does - **Configurable local word & link rules** — a starter English list and a few URL rules enforce on a fresh install. `/nmod test badword` verifies setup; review and extend the policy for your community. Configured terms catch leet-speak and spaced-out variants (`b4dw0rd`, `b a d w o r d`). - **Anti-spam** — message rate, duplicate/near-duplicate messages, excessive caps (player names exempt), character floods, and command spam. - **Progressive strikes** — repeated offences escalate automatically, and strikes decay over time. - **Censor or block** — replace matched words with `****` and keep the rest of the message, or block it outright. - **More than chat** — optionally moderate signs, books, anvil renames, and commands like `/msg`. - **Map-art scanning** — filled maps are checked for inappropriate images (optional cloud feature). - **Case history** — every detection is logged locally; browse it with `/nmod cases`. ## Trust & control - `/nmod test badword` — prove the bundled local rule works without executing an action. - `/nmod test <message>` — preview exactly how any message would be handled. - `/nmod doctor` — check local setup, account/usage access, credits, and the last known moderation-event result. - `/nmod mode monitor|enforce` — choose whether local detections alert or block; cloud has its own `/nmod cloudmode` setting. - `/nmod preset family|community|minimal` — apply a ready-made policy in one command. - `/nmod allow` — add exceptions that always win, to fix false positives instantly. - Staff alerts show who was flagged, on which surface, and what happened. ## Setup 1. Drop the jar in `plugins/` and restart. 2. Run `/nmod test badword`; expect **FLAGGED** and **blocked** in the preview. The test itself takes no action. Review the starter English list and actions in `config.yml`. 3. Optional cloud setup: [sign up](https://neomechanical.com/signup?src=neomoderation), create a key, run `/nmod setup <key>`, then `/nmod doctor` and `/nmod test hello`. 4. If `/nmod doctor` reports zero credits, use the [billing recovery page](https://neomechanical.com/billing?src=neomoderation_credits) and verify with `/nmod test hello`. 5. Review cloud detections in staff alerts, then run `/nmod cloudmode enforce` only when ready for cloud blocking. Use `/nmod mode monitor` if you want local rules to alert only. ## Commands `/nmod help` shows everything, grouped into Getting started, Rules & actions, Tools, and Admin. Aliases: `/neomod`, `/nmod`, `/neomoderation`. ## Privacy Local rules never leave your server. With a cloud key, checked messages are sent over HTTPS with no-store retention and no training use — run `/nmod privacy` to see exactly what is and isn't sent. ## Compatibility Supports Bukkit, Spigot, Paper, and Purpur on Minecraft 1.18.2 through the current Paper calendar releases (26.1, 26.2). Verified on Paper 1.18.2, 1.19.4, 1.20.6, 1.21.11, and 26.2 — including the supported 26.x line, not just the end-of-life 1.x line. The first four run in an automated server matrix; 26.2 is checked by hand.

# NeoModeration **Monitor-first chat moderation for Minecraft.** Runs configurable word/link matching and anti-spam locally with no account required; fresh installs only report what they would do until you explicitly enable enforcement. The bundled `badword` and `scam` entries are safe setup examples, not a complete profanity list. Add optional cloud moderation for context-aware checks and NSFW map-art scanning. Safe to try: **new installs start in monitor mode**, so you can see what *would* be actioned before anything blocks or punishes a player. Turn on enforcement when you're happy with the decisions. ## What it does - **Configurable local word & link rules** — the matching engine is ready immediately. The shipped safe examples let `/nmod test badword` verify setup; add the real terms and domains your community needs. Configured terms catch leet-speak and spaced-out variants (`b4dw0rd`, `b a d w o r d`). Blocking starts only after `/nmod mode enforce`. - **Anti-spam** — message rate, duplicate/near-duplicate messages, excessive caps (player names exempt), character floods, and command spam. - **Progressive strikes** — repeated offences escalate automatically, and strikes decay over time. - **Censor or block** — replace matched words with `****` and keep the rest of the message, or block it outright. - **More than chat** — optionally moderate signs, books, anvil renames, and commands like `/msg`. - **Map-art scanning** — filled maps are checked for inappropriate images (optional cloud feature). - **Case history** — every detection is logged locally; browse it with `/nmod cases`. ## Trust & control - `/nmod test badword` — prove the bundled local rule works without executing an action. - `/nmod test <message>` — preview exactly how any message would be handled. - `/nmod doctor` — check local setup, account/usage access, credits, and the last known moderation-event result. - `/nmod mode monitor|enforce` — evaluate safely, then switch on enforcement. - `/nmod preset family|community|minimal` — apply a ready-made policy in one command. - `/nmod allow` — add exceptions that always win, to fix false positives instantly. - Staff alerts show who was flagged, on which surface, and what happened. ## Setup 1. Drop the jar in `plugins/` and restart. 2. Run `/nmod test badword`; expect **FLAGGED** and **monitor alert only**. A fresh install still blocks nothing. 3. Optional cloud setup: [sign up](https://neomechanical.com/signup?src=neomoderation), create a key, run `/nmod setup <key>`, then `/nmod doctor` and `/nmod test hello`. 4. If `/nmod doctor` reports zero credits, use the [billing recovery page](https://neomechanical.com/billing?src=neomoderation_credits) and verify with `/nmod test hello`. 5. Review every enabled path, then run `/nmod mode enforce` only when ready. ## Commands `/nmod help` shows everything, grouped into Getting started, Rules & actions, Tools, and Admin. Aliases: `/neomod`, `/nmod`, `/neomoderation`. ## Privacy Local rules never leave your server. With a cloud key, checked messages are sent over HTTPS with no-store retention and no training use — run `/nmod privacy` to see exactly what is and isn't sent. ## Compatibility Supports Bukkit, Spigot, Paper, and Purpur on Minecraft 1.18.2 through the current Paper calendar releases (26.1, 26.2). Verified on Paper 1.18.2, 1.19.4, 1.20.6, 1.21.11, and 26.2 — including the supported 26.x line, not just the end-of-life 1.x line. The first four run in an automated server matrix; 26.2 is checked by hand.

# NeoModeration **Monitor-first chat moderation for Minecraft.** Runs configurable word/link matching and anti-spam locally with no account required; fresh installs only report what they would do until you explicitly enable enforcement. The bundled `badword` and `scam` entries are safe setup examples, not a complete profanity list. Add optional cloud moderation for context-aware checks and NSFW map-art scanning. Safe to try: **new installs start in monitor mode**, so you can see what *would* be actioned before anything blocks or punishes a player. Turn on enforcement when you're happy with the decisions. ## What it does - **Configurable local word & link rules** — the matching engine is ready immediately. The shipped safe examples let `/nmod test badword` verify setup; add the real terms and domains your community needs. Configured terms catch leet-speak and spaced-out variants (`b4dw0rd`, `b a d w o r d`). Blocking starts only after `/nmod mode enforce`. - **Anti-spam** — message rate, duplicate/near-duplicate messages, excessive caps (player names exempt), character floods, and command spam. - **Progressive strikes** — repeated offences escalate automatically, and strikes decay over time. - **Censor or block** — replace matched words with `****` and keep the rest of the message, or block it outright. - **More than chat** — optionally moderate signs, books, anvil renames, and commands like `/msg`. - **Map-art scanning** — filled maps are checked for inappropriate images (optional cloud feature). - **Case history** — every detection is logged locally; browse it with `/nmod cases`. ## Trust & control - `/nmod test badword` — prove the bundled local rule works without executing an action. - `/nmod test <message>` — preview exactly how any message would be handled. - `/nmod doctor` — check local setup, account/usage access, credits, and the last known moderation-event result. - `/nmod mode monitor|enforce` — evaluate safely, then switch on enforcement. - `/nmod preset family|community|minimal` — apply a ready-made policy in one command. - `/nmod allow` — add exceptions that always win, to fix false positives instantly. - Staff alerts show who was flagged, on which surface, and what happened. ## Setup 1. Drop the jar in `plugins/` and restart. 2. Run `/nmod test badword`; expect **FLAGGED** and **monitor alert only**. A fresh install still blocks nothing. 3. Optional cloud setup: [sign up](https://neomechanical.com/signup?src=neomoderation), create a key, run `/nmod setup <key>`, then `/nmod doctor` and `/nmod test hello`. 4. If `/nmod doctor` reports zero credits, use the [billing recovery page](https://neomechanical.com/billing?src=neomoderation_credits) and verify with `/nmod test hello`. 5. Review every enabled path, then run `/nmod mode enforce` only when ready. ## Commands `/nmod help` shows everything, grouped into Getting started, Rules & actions, Tools, and Admin. Aliases: `/neomod`, `/nmod`, `/neomoderation`. ## Privacy Local rules never leave your server. With a cloud key, checked messages are sent over HTTPS with no-store retention and no training use — run `/nmod privacy` to see exactly what is and isn't sent. ## Compatibility Supports Bukkit, Spigot, Paper, and Purpur on Minecraft 1.18.2 through the current Paper calendar releases (26.1, 26.2). Verified on Paper 1.18.2, 1.19.4, 1.20.6, 1.21.11, and 26.2 — including the supported 26.x line, not just the end-of-life 1.x line. The first four run in an automated server matrix; 26.2 is checked by hand.

# NeoModeration **Monitor-first chat moderation for Minecraft.** Runs configurable word/link matching and anti-spam locally with no account required; fresh installs only report what they would do until you explicitly enable enforcement. The bundled `badword` and `scam` entries are safe setup examples, not a complete profanity list. Add optional cloud moderation for context-aware checks and NSFW map-art scanning. Safe to try: **new installs start in monitor mode**, so you can see what *would* be actioned before anything blocks or punishes a player. Turn on enforcement when you're happy with the decisions. ## What it does - **Configurable local word & link rules** — the matching engine is ready immediately. The shipped safe examples let `/nmod test badword` verify setup; add the real terms and domains your community needs. Configured terms catch leet-speak and spaced-out variants (`b4dw0rd`, `b a d w o r d`). Blocking starts only after `/nmod mode enforce`. - **Anti-spam** — message rate, duplicate/near-duplicate messages, excessive caps (player names exempt), character floods, and command spam. - **Progressive strikes** — repeated offences escalate automatically, and strikes decay over time. - **Censor or block** — replace matched words with `****` and keep the rest of the message, or block it outright. - **More than chat** — optionally moderate signs, books, anvil renames, and commands like `/msg`. - **Map-art scanning** — filled maps are checked for inappropriate images (optional cloud feature). - **Case history** — every detection is logged locally; browse it with `/nmod cases`. ## Trust & control - `/nmod test badword` — prove the bundled local rule works without executing an action. - `/nmod test <message>` — preview exactly how any message would be handled. - `/nmod doctor` — check local setup, account/usage access, credits, and the last known moderation-event result. - `/nmod mode monitor|enforce` — evaluate safely, then switch on enforcement. - `/nmod preset family|community|minimal` — apply a ready-made policy in one command. - `/nmod allow` — add exceptions that always win, to fix false positives instantly. - Staff alerts show who was flagged, on which surface, and what happened. ## Setup 1. Drop the jar in `plugins/` and restart. 2. Run `/nmod test badword`; expect **FLAGGED** and **monitor alert only**. A fresh install still blocks nothing. 3. Optional cloud setup: [sign up](https://neomechanical.com/signup?src=neomoderation), create a key, run `/nmod setup <key>`, then `/nmod doctor` and `/nmod test hello`. 4. If `/nmod doctor` reports zero credits, use the [billing recovery page](https://neomechanical.com/billing?src=neomoderation_credits) and verify with `/nmod test hello`. 5. Review every enabled path, then run `/nmod mode enforce` only when ready. ## Commands `/nmod help` shows everything, grouped into Getting started, Rules & actions, Tools, and Admin. Aliases: `/neomod`, `/nmod`, `/neomoderation`. ## Privacy Local rules never leave your server. With a cloud key, checked messages are sent over HTTPS with no-store retention and no training use — run `/nmod privacy` to see exactly what is and isn't sent. ## Compatibility Supports Bukkit, Spigot, Paper, and Purpur on Minecraft 1.18.2 through the current Paper calendar releases (26.1, 26.2). Verified on Paper 1.18.2, 1.19.4, 1.20.6, 1.21.11, and 26.2 — including the supported 26.x line, not just the end-of-life 1.x line. The first four run in an automated server matrix; 26.2 is checked by hand.

# NeoModeration 1.4.1 — Activation clarity NeoModeration 1.4.1 makes the first-run and cloud-recovery paths explicit. A fresh install still starts safely in **monitor mode**: it detects and reports, but does not block or punish until an admin runs `/nmod mode enforce`. ## Improved - **Guided first test.** Startup, help, README, and marketplace copy now point to `/nmod test badword`. It proves the bundled local rule works and states that monitor mode will not execute an action. - **Truthful key setup.** Saving a key no longer claims cloud moderation is active before it has been checked. `/nmod doctor` verifies the account and usage API, latency, and the current credit balance; `/nmod test hello` verifies the moderation-events path separately. - **Actionable cloud recovery.** HTTP 401/403 is reported as a rejected key, while HTTP 402 is reported separately as insufficient credits. Recovery commands provide the matching key or billing link without echoing an API key. - **Honest billing classification.** Ordinary Minecraft chat uses the platform's `general` content class. Only NSFW map-art image scanning declares `nsfw`. - **Separated cloud health.** Chat tests and map-art scans feed the moderation-event circuit/status state. Account checks cannot make a broken events endpoint look healthy. ## Compatibility - No configuration migration is required. - Existing enforce-mode servers remain in enforce mode. Monitor is only the default for a genuinely new config. - Supports Minecraft 1.18.2–1.21.11 on Bukkit, Paper, Purpur, and Spigot. Folia is not advertised. - Local word, URL, and anti-spam rules continue to run when cloud moderation is unavailable or out of credits. ## Verified - Automated regression coverage includes 401/403, 402, other 4xx, transient failures, separated account/event health, recovery routing, locale parity, billing classes, and the guided monitor-mode test path. - The release matrix checks Paper 1.18.2, 1.19.4, 1.20.6, and 1.21.11.

# NeoModeration **Automatic chat moderation for Minecraft.** Blocks swearing, spam, advertising links, and inappropriate map art — instantly, with local rules that need no account or setup. Add optional cloud moderation when you want smarter, context-aware checks. Safe to try: **new installs start in monitor mode**, so you can see what *would* be actioned before anything blocks or punishes a player. Turn on enforcement when you're happy with the decisions. ## What it does - **Local word & link filtering** — works the instant you install it. Leet-speak and spaced-out bypasses are caught (`b4dw0rd`, `b a d w o r d`). - **Anti-spam** — message rate, duplicate/near-duplicate messages, excessive caps (player names exempt), character floods, and command spam. - **Progressive strikes** — repeated offences escalate automatically, and strikes decay over time. - **Censor or block** — replace matched words with `****` and keep the rest of the message, or block it outright. - **More than chat** — optionally moderate signs, books, anvil renames, and commands like `/msg`. - **Map-art scanning** — filled maps are checked for inappropriate images (optional cloud feature). - **Case history** — every detection is logged locally; browse it with `/nmod cases`. ## Trust & control - `/nmod test <message>` — preview exactly how any message would be handled. - `/nmod doctor` — one command checks your whole setup and connectivity. - `/nmod mode monitor|enforce` — evaluate safely, then switch on enforcement. - `/nmod preset family|community|minimal` — apply a ready-made policy in one command. - `/nmod allow` — add exceptions that always win, to fix false positives instantly. - Staff alerts show who was flagged, on which surface, and what happened. ## Setup 1. Drop the jar in `plugins/` and restart. 2. Local rules work immediately (in monitor mode on a fresh install). 3. Optional: run `/nmod setup <key>` for cloud moderation, then `/nmod doctor` to confirm. 4. Happy with it? `/nmod mode enforce`. ## Commands `/nmod help` shows everything, grouped into Getting started, Rules & actions, Tools, and Admin. Aliases: `/neomod`, `/nmod`, `/neomoderation`. ## Privacy Local rules never leave your server. With a cloud key, checked messages are sent over HTTPS with no-store retention and no training use — run `/nmod privacy` to see exactly what is and isn't sent. ## Compatibility Works on Bukkit, Spigot, Paper, Purpur, and Folia-friendly forks, Minecraft 1.13 and newer. Verified on Paper 1.18.2, 1.19.4, 1.20.6, and 1.21.11.