One bot instead of ten. AI that remembers your members, moderation that escalates on its own, tickets that ask the follow-up questions for you, and a verify gate that keeps ban evaders out. Over 200 slash commands, every one of them configured from the dashboard.
Seven areas nearly every server needs on day one. The full directory with every module sits further down.
Mrs. Maid is deliberately big. Instead of stitching ten separate bots together, everything runs in one process: same database, same audit log, same dashboard. Changes apply on the next event tick.
Most of the damage on a Discord server comes from someone who was already removed once. Mrs. Maid checks every new member in the browser, turns what it finds into one number, and only pulls your staff in when that number is not obvious.
They get one button. Nothing to install, no command to remember.
They agree, sign in with Discord, and the page reads how their device and browser behave. The raw IP is not kept.
Device matches, join timing, account age and username patterns become a single number between 0 and 100.
A clean member is through in seconds. Everyone else goes into a staff queue with a case file to read.
The gate ships with the bot. Pick the channel, pick the role a verified member gets, and set the score below which someone goes to review instead of straight through.
Both run on a bearer key you create in the dashboard, both have a free tier with a daily quota, and both are documented endpoint by endpoint.
Curated reaction GIFs for your own bot. One request gives you one URL you can drop straight into a Discord embed. SFW and NSFW stay apart.
curl -H "Authorization: Bearer mm_gif_..." https://mrs-maid.cc/api/v1/sfw/hug
You start a session, the member walks through the verify page, you get the verdict. Pushed to your webhook, or polled if you missed it.
POST https://mrs-maid.cc/api/verify/v1/session
Join the official Mrs. Maid server and try everything in a real environment. Mod logs, AI replies, tickets, self-roles. All under real conditions.
No. This dashboard talks to the same bot you already use. Sign in with Discord and we'll list every guild where you can manage server.
Encrypted with the bot's master key. The dashboard never holds them in plain text and never sends them to your browser.
Only to Discord, and only during sign-in. We fetch your guilds, then revoke the access token immediately.
Mammouth is a European multi-model gateway with a clear privacy stance: no storage of prompts or responses for training, no hand-off to third parties. We route requests through their gateway to models like Grok, Claude, GPT-4 or Mistral. You pay one provider instead of five, and the bot stays in a sovereign, GDPR-compliant setup. More: mammouth.ai.
Per request the bot assembles a short context (the last few messages + the system prompt from your server settings) and sends it to Mammouth over HTTPS. We never persist the full message content anywhere. Behavior analysis reads text in flight, generates a summary, and discards the originals. Stats are pure numbers: message count, voice minutes, XP. Content never enters the database.
According to Mammouth: zero retention on API requests (prompts are NOT stored, NOT used for training, NOT shared with the model vendors). Hosted in Europe, GDPR-compliant. Full details in their privacy policy: mammouth.ai/privacy. If your server does not want AI at all, disable it in the dashboard (casual probability 0, no AI channel, no API key set = AI completely off).
Yes. The GIF API serves curated reaction GIFs for your own bot, the Verify API checks accounts for alts and returns a trust score. Both need a key from the dashboard and both have a free tier. Details are in the GIF API docs and the Verify API docs.
No. The scan turns device and browser signals into a fingerprint and keeps that. The raw IP is never written down. What your staff gets is a score and the list of signals behind it, and the full detail is in the privacy policy.
That case does not get kicked. A device match on its own only sends someone to review, where a moderator sees the score, the signals and a downloadable case file, then approves or denies with one click. The only thing that acts without a human is a burst of joins from one device, and that is not a pattern a single person produces by accident.