Skip to content

For Minecraft server owners

Keep your server alive: see who's slipping away, nudge them back the right way, and know what actually worked.

Minealyze spots the players spacing out their visits, drafts a short message to bring them back (never pay-to-win), then shows you, numbers in hand, whether it worked.

Free to start · No credit card · Plugin in 5 minutes

emberfall.mc / overviewLive · 2 min ago

Still playing after 7 days · newcomers

13%

Store · 7 days

−8% vs last week

€312.40

32 players supported the store

Who's slipping away

24 slipping
  • ZylkorSession-cadence break
  • PandaRougePlaytime dropping
  • MirayaLast purchase long ago

☀️ Morning recap sent

Discord · 08:07

Demo · synthetic data
Read the walkthrough as text
  1. 3 of 24 flagged players are slipping away: Zylkor, PandaRouge, Miraya.
  2. Still playing after 7 days: 13% of newcomers, redrawn every few seconds as the engine rescans. The number itself does not move.
  3. The store follows the server: €312.40 over 7 days, down 8% from the week before.
  4. A morning recap went out on Discord at 08:07.

Built by server owners, for server owners. We know the numbers that keep a server alive.

What we read, what we never read

How it works

Understand. Act. Prove it worked.

Three steps, one thread: your players slipping away, a nudge you confirm, the proof by day 14.

  1. Understand

    The Who's slipping away list shows you who, and why.

    The plugin measures sessions, read-only. Every flagged player comes with their top reason, in plain words: "Not seen in 9 days (usually plays every 2)."

    • The top reason, in plain words, with real units
    • Sorted by urgency, never by money
    • Measured from sessions, never from chat
    emberfall.mc / who's slipping away

    > reading sessions…

    24 players slipping · 🔒 19 on a higher plan
    Demo · synthetic data
    Show the walkthrough as text
    1. The plugin reads the sessions.
    2. Five players arrive, each with their top reason in plain words.
    3. Zylkor moves to the top: his rhythm broke 6 days ago.
    4. 24 players slipping in total, 19 names locked on this plan.

    Measured from sessions, never inferred from chat.

  2. Act

    A nudge is previewed, then confirmed. Never the other way round.

    You see everything before it sends: the message, the reward (never pay-to-win), the do-not-disturb players excluded. A few players are set aside, so you can compare later.

    • "No gameplay advantage": you attest it yourself
    • A few players set aside, to compare later
    • Everything goes in the log
    emberfall.mc / nudge

    Confirm the nudge

    players slipping away · 24 targeted
    target
    24
    nudged
    21
    set aside
    3
    do-not-disturb excluded
    1

    MESSAGE AI-written

    « A gift is waiting for you: 1 Ember Key, come back and claim it! »

    21 players will receive the key

    Only this button sends. Written to the log, server-side.

    Demo · synthetic data
    Show the walkthrough as text
    1. The preview is ready: 21 nudged, 3 set aside to compare, 1 do-not-disturb excluded.
    2. You attest the reward gives no gameplay advantage. The AI cannot check this box for you.
    3. You confirm. It is the only button that sends.
    4. In-game: the key arrives in the chat.
    5. Logged at 09:03: 21 nudged, 3 set aside. The proof arrives at day 14.

    Nothing goes out before you confirm.

  3. Prove

    By day 14, the effect is measured, not narrated.

    12 of the 21 nudged players came back, versus 1 of the 3 set aside: about +24 points. Small group? The page says so.

    • Nudged players compared to players set aside
    • Players first, the percentage second
    • Not enough data? We say so, we don't invent
    emberfall.mc / proof
    Nudged (21)Set aside (3)
    +24 ptsrange [+6 ; +41]small groupwide range

    21 nudged · 3 set aside · day 14 check

    Demo · synthetic data
    Show the walkthrough as text
    1. Axes are set: 21 nudged players versus 3 set aside.
    2. Both curves are drawn from day 0 to day 14.
    3. Day 14: 12 of the 21 nudged players came back, versus 1 of the 3 set aside.
    4. Real effect: about +24 points, likely between +6 and +41. Small group, so the range is wide, and we say so.

    When the group is small, we write it.

The thread

Zylkor is slipping away. Here's how he comes back.

One player, end to end: spotted, nudged the right way, and the proof he did not come back by chance.

  1. 01 · THE SIGNAL

    His playtime drops 63% in 14 days

    The plugin measures sessions: Zylkor used to play every 2 days, nothing for 6 days now. Measured from his sessions, never inferred from chat.

  2. 02 · WHO'S SLIPPING AWAY

    Flagged before he's gone for good

    Verdict: fading fast. Top reason: his session rhythm broke. The list puts him on top while he can still be brought back, sorted by urgency, never by money.

  3. 03 · THE NUDGE

    Preview, then confirm. Never the other way round.

    21 nudged, 3 set aside to compare, 1 do-not-disturb excluded upfront. Nothing is sent without your confirmation, and you attest it yourself: no gameplay advantage.

  4. 04 · IN GAME

    An Ember Key waits at his next login

    The command runs server-side when Zylkor reconnects. Everything goes in the log: you know what was sent, when, to whom.

  5. 05 · THE PROOF

    What the nudge actually changed, measured honestly

    Day 14: 12 of the 21 nudged players came back, versus 1 of the 3 set aside. That points to about +24 points, but 3 players set aside is too few to settle it, and the page says exactly that. To get an answer you can trust, set more players aside next time.

The 3 players set aside make the proof solid: without them, +24 points would be an anecdote. Emberfall is a fictional server; the flow is the real product.

Take over in the product tour

An owner's week

The server comes to you, not the other way round.

Maya runs Emberfall from her phone. Every morning at 08:07, the recap lands in her Discord: 3 things to do, or "nothing urgent today".

Maya

Owner of Emberfall (demo server)

  • 08:07 · The morning recap lands in Discord

    2 players slipping away (1 of them a pillar) · emeralds alert ×4.2 at 03:40

    Measured · from her data
  • 08:09 · She rereads the drafted message, confirms the nudge

    21 nudged · 3 set aside to compare · 1 do-not-disturb excluded

    Confirmed · in the log
Nothing left without her: every send was confirmed by Maya, and it is all in the log.

Day to day

The signal shows up while there's still time.

Your newcomers' first 5 minutes, step by step. The morning recap in your Discord at 08:07. An alert at 03:40 when a number jumps out of the ordinary. Not three weeks later in a CSV export.

emberfall.mc / first 5 minutes
Demo · synthetic data
Show the walkthrough as text
  1. 412 newcomers arrive that week; each step counts how many keep going.
  2. Step 2, tutorial done: 86%, 354 players.
  3. Step 3, first chat: 54%, 223 players. The break from step 2 is highlighted: minus 32 points.
  4. Step 4, first chest opened: 38%, 157 players.
  5. Step 5, back the next day: 24%, 99 players. Tip: schedule a social event right after the tutorial.

Players are counted first; percentages are derived from the real counts, never typed by hand.

emberfall.mc / morning recap
Minealyze Bot

Morning recap · Emberfall

SLIPPING AWAY

24

+5 vs prev. 7d

STORE 7D

€312.40

-8.1% vs prev. 7d

ALERT

×4.2

Demo · synthetic data
Show the walkthrough as text
  1. The message arrives at 08:07, before you open anything.
  2. Players slipping away: 24, 5 of them new this week.
  3. The store follows: €312.40 over 7 days, down 8.1% vs the week before.
  4. One alert: emeralds per hour at 4.2 times the usual. The link opens the list.
emberfall.mc / alerts
EMERALDS / HOURHigh

7-day emeralds-per-hour curve, spike at 4.2 times the usual level, spotted at 03:40.

×4.2

Lead: an unregulated automated farm

usual level, 7d310/h · peak1302/h

measuredProjected shop sales: up
Demo · synthetic data
Show the walkthrough as text
  1. The 7-day curve draws in: a steady rhythm, day and night cycles.
  2. A spike appears at 03:40: 4.2 times the usual level.
  3. An alert card opens. Lead: an unregulated automated farm. You decide, nothing acts on its own.
  4. Severity: high.
  5. Measured impact: projected shop sales are up.

The morning recap and the alerts: both land in your Discord.

The nevers

What Minealyze refuses to do.

  • Never pay-to-win

    Rewards stay cosmetic or social, never a gameplay advantage. That is Mojang’s rule, and ours.

  • Nothing sends without you

    Preview, confirm, log: Minealyze proposes, you decide. And it never bans or punishes a player.

  • Players set aside, to prove it

    If we cannot prove a result, we write "not enough data". Never a fake number to keep you around.

  • Do not disturb, respected

    A player who said no never gets another message. And caps prevent over-messaging: too many nudges drives people away.

  • Your players never feed an AI

    Usernames are coded before any AI call, chat is never collected by default, and your data is never sold.

  • Kill switch, always

    The kill switch stops auto-nudges immediately, whatever your plan.

The copilot

Ask your question. The numbers, though, are measured.

The copilot reads your data and writes the answer; it never computes a number itself. If it doesn't have the number, it doesn't have an opinion.

  • Every number shown comes from a measurement, displayed right next to it
  • It can propose a nudge; you preview and you confirm
  • Question meter and outages shown as they are
And when it can't answer?
emberfall.mc / copilot

« Who is slipping away this week? »

> reading your session data…

3/25 this month
Demo · synthetic data
Show the walkthrough as text
  1. The question is typed: who is slipping away this week?
  2. The copilot reads your data: real measurements, nothing invented.
  3. The answer is written; the numbers next to it are shown as measured.
  4. Every number is measured by Minealyze; the AI only writes the sentence.
  5. The question meter is re-read after the answer: 4 out of 25 this month.
  6. A proposal appears: nudge these 5 players. Preview only, never sent automatically.

The measurement is shown next to the sentence, as-is. The numbers are computed by Minealyze; the AI only puts them into words. Everything is logged.

Not another dashboard

A stats page tells you what happened. Minealyze tells you what to do, then proves it worked.

A dashboard you have to open every day to notice anything

A recap that lands in your Discord at 08:07, nothing to open to know if it's fine

A risk score with no explanation

Every player comes with a top reason, in plain words, measured from their sessions

An alert, then nothing, and you're on your own to act

Preview → confirm → log, then the real effect measured: nudged players vs players set aside

Where the data comes from

A plugin, not a mod. Measurements, not chat.

Four verifiable steps, from your server to your Discord, and what each one refuses to read.

  1. 01 · Source

    Paper plugin

    join · quit · playtime

    Read-only. Zero chat messages read.
  2. 02 · Source

    Your shop (optional)

    purchases · Tebex webhook or Azuriom API

    Payments stay with your shop.
  3. 03 · Compute

    Engine

    player groups · alerts

    Computed server-side, data in the EU. No AI decides anything.
  4. 04 · Delivery

    Your dashboard / Discord

    who's slipping away · 08:07 recap

    "Do not disturb" honored, always.

4 steps · 0 chat messages read · 0 client mods · EU data

The GDPR details, in plain words

Install

Three steps. Five minutes. No hidden config.

  1. Drop the plugin

    Put minealyze-collector.jar in /plugins and restart. Read-only: it never writes to your world.

    plugins/minealyze-collector.jar
  2. Paste the server key

    Paste the key into the plugin config. Your shop, Tebex or Azuriom, is optional and can be linked later.

    plugins/Minealyze/config.yml
  3. Get your first morning recap

    It lands in your Discord the next day, at 08:07.

Questions

The questions we get before people buy.

Real doubts from server owners, answered plainly.

Does this replace a moderator?

No. Minealyze measures and flags; decisions (banning, refunding, messaging) stay yours.

Will my players know they are being watched?

Nobody reads anyone. Minealyze only measures game events (join, quit, purchase), never chat. And a player can say "do not disturb": it is respected, immediately.

Which shop plugin does it work with?

Tebex, via a signed webhook, and Azuriom, which we query periodically with a read-only token since it sends no webhook. The shop is optional either way: Minealyze works without one.

How long before I see something useful?

The first morning recap arrives the day after you connect the plugin. No history needed.

Does the copilot make up numbers?

No. Every number it cites is computed by Minealyze from your data; the AI only puts it into words. If it doesn't have the number, it says so.

Can I export and delete my data?

Yes, any time, from settings: CSV export and full deletion.

More questions? Write to us

Next time someone starts slipping, you'll see it coming.

Connect your server. While your data comes in, you explore the Emberfall demo, clearly marked as a demo: the screen is never empty.

Connect my server