RetentionCore follows every new player's first hours, names the reason each one quit, and steps in while they are still online.
Paper, Purpur, Spigot, Folia and Velocity networks · Minecraft 1.20.6 to 26.3 · no port to open
Of … new players, … came back the next day.
A week after a player's last visit, RetentionCore looks at how that session ended and names the cause: died to lava and lost 20 stacks, asked a question nobody answered, typed commands that don't exist, lag, a kick from the anti-cheat. Thirteen reasons, each with a confidence score and the last minutes of play as a timeline.
Give every server list and ad its own subdomain. RetentionCore compares how many players each one brings, how many are still around after a week, what they spend in your store and what each staying player cost you.
Every lost player's last position goes on a map, and clusters are found for you. A ring around a lava canyon next to your /rtp zone, or around the spawn tutorial, tells you exactly where to look.
No setup: every first chat, command, world and milestone is recorded automatically. RetentionCore then shows which steps players who stay almost always take, and players who leave almost never do, so you know what your tutorial should teach.
When a new player asks something in chat and nobody answers within 30 seconds, your staff gets pinged in game and on Discord. A player who loses their items gets pointed to /back. A typo gets a "did you mean /warp?". Each action is yours to configure.
To prove it works, a small share of situations is left alone on purpose. The panel compares how many helped and not-helped players come back, with a real significance test.
Split new players between two starter kits, spawns or tutorials, and see which keeps more of them. Variants are exposed as permissions and PlaceholderAPI placeholders.
Every Monday: retention, what changed, the biggest problem and a concrete suggestion, written in plain words.
Invite admins and moderators with a link. They log in with Discord and see the dashboard, with viewer or admin rights.
A 0 to 100 score for every online player, also as %retentioncore_risk% for your scoreboard or staff tools.
One jar for the proxy and every backend. Each backend is its own server in the panel, with the real join address forwarded by the proxy.
A Discord message when a server stops sending data, and another when it is back.
The plugin keeps all raw data on your server. The panel only receives finished reports, and you can use the built-in dashboard without the panel at all.
Drop the jar into plugins and restart. On a Velocity network, put the same jar on the proxy and on every backend. Tracking starts right away.
Log in with Discord and give the server a name. The panel shows a link command with your key.
rc cloud link <key> in the console. The plugin only connects outward over HTTPS, so there is no port to open on your host.
Paper, Purpur and Spigot from 1.20.6 up to 26.3, Folia, and Velocity proxies. Deprecated server events are isolated, so a future server version can switch off a single signal but not the plugin.
Yes. Put the same jar on the Velocity proxy and on every backend server. The backends do all the tracking and reporting; the proxy part only tells them which address each player joined with, so source tracking keeps working behind the proxy. BungeeCord backends are tracked too, without the join address.
Finished reports (percentages, counts, the last 25 players who left) and the live view. Player details are fetched from your server only when someone on your team opens them. Chat messages and IP addresses are never stored.
No. Every database query and upload runs on its own background thread, and events only update small counters in memory.
Yes. /rc simulate 1000 fills a test server with simulated players so every report has data, and /rc clearsim removes them again.