Source: https://respawn.sh/ # The agentic ops layer for game servers An operator that never sleeps, backups you can read like a diff, and analytics that tell you why players come and go. Get started free (https://auth.respawn.sh/signup) Book a demo (https://respawn.sh/book) Music and sound effects, no speech. On screen: respawn.sh, the agentic ops layer for game servers, and its three products: Veyra runs your server, Ashryn brings it back, Scrye sees it coming. ## Veyra Runs your server with you. ### Fixes what breaks, at the level you choose Pick how much Veyra does on its own. Every fix you approve runs with a snapshot first, 15 minutes of watching and an automatic undo if things get worse. Veyra investigates problems and suggests fixes. Every fix waits for your approval. NotificationsAutopilot: Notify Survival is crash looping02:16 9 crashes in 40 minutes, each a NoSuchMethodError in EssentialsX. It started 6 minutes after the 02:08 plugin update. Sent to the notification center and email. Veyra took no action. Ask Veyra about thisOpen the console Waiting for approvalAutopilot: Suggest AI diagnosisHigh confidence EssentialsX 2.21.0 calls a method Paper 1.21.4 build 212 doesn't have. It throws the moment a player joins. Veyra wants to: roll back a pluginWaiting for approval Roll EssentialsX back from 2.21.0 to 2.20.1 and restart Survival. A checkpoint snapshot first, 15 minutes of checks, automatic rollback if the server gets worse. ApproveDecline Server timelineAutopilot: Protect Took a backup of Survival 02:16 · restore point before any changeDone on its own Opened an incident: crash loop 02:16 · 9 crashes in 40 minutesDone on its own Added notes to the incident 02:17 · EssentialsX 2.21.0 named as the likely causeDone on its own Veyra wants to: roll back a pluginWaiting for approval Roll EssentialsX back from 2.21.0 to 2.20.1 and restart Survival. A checkpoint snapshot first, 15 minutes of checks, automatic rollback if the server gets worse. ApproveDecline ### A crash, explained and fixed The cause, the evidence and a fix you approve. Survival crashed at 02:14 NoSuchMethodError in EssentialsX, 9 crashes in 40 minutesFixed Fixed this way on 37+ servers: 37 Paper servers hit this after updating to 2.21.0; rolling back fixed 35 of them. AI diagnosisHigh confidence EssentialsX 2.21.0, installed by the 02:08 update, calls Player.getRespawnLocation(), which Paper 1.21.4 build 212 doesn't have. It throws the moment a player joins. Root cause: Plugin built for a newer Paper build Why we think so - The crash starts 6 minutes after EssentialsX moved from 2.20.1 to 2.21.0. (2 sources) - Every crash is a NoSuchMethodError inside EssentialsX's join handler. (1 source) - The same fix worked on this server on October 12. (1 source) Proposed fix Roll back to 2.20.1. 2.20.1 ran for 19 days on this Paper build without a crash. Ask Veyra about thisWas this right?HelpfulWrong EssentialsX is back on 2.20.1 and the server restarted. Undo: go back to EssentialsX 2.21.0Watching for 15 minutes, 11 left ### Remembers what worked Fixes, preferences and known issues, kept for each server. OrganizationsSakura NetworkSurvivalVeyra ### VeyraSurvival Ask about this server, review what it proposes, and choose what it may do on its own. ChatJobsApprovals and activityAutopilotMemoryHealth and safety #### What Veyra remembers about Survival Short facts and fixes it learned here. Remove anything that's wrong or out of date. At a glance What it's like Paper 1.21.4 survival server behind Velocity, 38 plugins Preferences Restarts 04:00 to 06:00 Tokyo time with a 5 minute warning Lately Rolled EssentialsX back to 2.20.1 after a join crash Fix that workedSaved Oct 12, last used 20 min ago EssentialsX 2.21.x breaks on Paper build 212: roll back to 2.20.1 until Paper 213. EditForget PreferenceSaved Sep 27, last used 1 h ago Restarts only between 04:00 and 06:00 Tokyo time, and warn players 5 minutes before. EditForget Known issueSaved Oct 27, last used 2 days ago Dynmap full renders push tick time past 60 ms; pause renders during events. EditForget PreferenceSaved Oct 30, last used 3 days ago Ask before changing anything players can see. EditForget ### Ask it anything about your server Growth trends, top regions, the hostnames players join from and your network topology. Why did players drop on Friday night?↑ Players fell 38% after 21:00. Most of the drop came from the EU proxy, where tick time reached 64 ms while Dynmap rendered. I can pause renders at peak hours. From Players, tick time, plugin timings, console ## Ashryn Every world, always restorable. ### Every backup, at a glance Encrypted and incremental, so a long history stays small. OrganizationsSakura NetworkSurvivalBackups Ashryn ### BackupsSurvival 1,342 backups stored for Survival Back up nowRestoreSettings Storage 38 GB First backup 9.8 GB , changes since 29 GB Used by this server Latest backup 28 min ago Nov 6, 8:12 AM Backup history Each square is a day. Pick one to jump the list there. 159 backups are older than 30 days and added 19 GB. Old history rarely needs every backup. Keeping a few per day frees storage.Review them MayJunJulAugSepOctNov Fewer More ### Compare any two backups, then open just what you need Which files and plugins changed, and when. Look inside any backup and bring back single files without a full restore. Backup at 06:00 compared with yesterday+ plugins/LuckPerms-5.4.145.jar - plugins/LuckPerms-5.4.141.jar + world_nether: 1,204 chunks changed server.properties: unchanged Scanned. Nothing unexpected. ### Back in a minute or two A safety snapshot first, then 15 minutes of watching. OrganizationsSakura NetworkSurvivalBackups Ashryn ### BackupsSurvival 1,342 backups stored for Survival Back up nowRestoreSettings Restoring Survival, 62% done.Cancel restore Safety snapshot taken first. Applies before worlds load, then 15 minutes of watching. Storage 38 GB First backup 9.8 GB , changes since 29 GB Used by this server Latest backup 35 min ago Nov 6, 8:12 AM ## Scrye Knows what normal looks like. ### Where your players join from Regions, hostnames and growth. ### Why it slowed down OrganizationsSakura NetworkScryeSurvivalSlowdowns Scrye ### SlowdownsSurvival Find the change most likely behind a slowdown. Compares the last healthy backup with the first one after the slowdown, and how much time each plugin took before and after it began. When did the slowdown start? Latest incident1 hour agoThis morningYesterdayPick a time Most likely cause Comparing the backups from Nov 6, 5:12 AM (healthy) and Nov 6, 5:42 AM (slow), and plugin time before and after. dynmap was updated, then got slower Strong match Dynmap-3.7.jar → Dynmap-3.8.jar, handler time 40 to 168 ms/min, worst call 20 to 61 ms Download the previous version from the healthy backupRestore the healthy backup ### Normal for this hour What changed in the last hour - ▴ Entities went up 36% (3,050 to 4,147) during the last hour. - ▾ Players (this world) went down 28% (49 to 36) during the last hour. Normal for this hour Each reading against the same hour of the week over the last five weeks (Asia/Tokyo). PlayersUnusually low 17 Normal for Friday 20:00: 34 to 52 Players 60% below a normal Friday at 20:00 TPSNormal 19.9 Normal for Friday 20:00: 19.6 to 20 TPS normal for a Friday evening p95 tick timeNormal 31 ms Normal for Friday 20:00: 26 ms to 41 ms Tick time normal for this hour MemoryNormal 61% Normal for Friday 20:00: 55% to 72% Memory normal for this hour Target 99.9% uptime over 30 daysOn track 50% of the error budget left Half the month's downtime budget left: 21 of 43 minutes used ### Alerts that reach you Discord and email, only when something is unusual for this hour. ## Built for the servers you run Minecraft Hytale FiveM ## Hosts Give it to every customer, under your brand. ### A respawn.sh tab inside your panel Pterodactyl, Pelican and Calagopus, in your panel's own theme. OverviewConsoleFilesBackupsAnalyticsAlerts2AssistantActivityUsage RunningPaper 1.21.4, Java 21.0.4respawn.sh plugin v2.14.0Update v2.15.0 ready, applies on the next restartRestart Right now Trend: last 24 hours TPS 17.6 Watchunder 18.5 Tick time (p95) 69.8 ms Criticalover 50 ms CPU 66.0% Watchusually 39.9% Heap used 2.7 GB usually 2.6 GB Players online 14 usually 9 Entities 3,976 usually 2,596 Recent incidentsAll incidents → WatchEntities rose to 4,180, well above normal4 h ago CriticalTick time (p95) rose to 74 ms, far above normal5 h ago ResolvedHeap used climbed for 40 minutes2 days ago BackupsAll backups → Next backupIn 5 hLatest1 h agoSpace used19.9 GB ### Customers restore their own backups Fewer tickets for you. ConsoleFilesBackups 19 h ago Full, 14.7 GB.Restore Yesterday Incremental, 980 MB.Restore Yesterday Safety backup, 691 MB.Undo 2 days ago Incremental, 784 MB.Restore Find a backup from before a time 1 hour agoThis morningYesterdayExact time 13 h ago is the last backup before this morning, 06:00.Restore it Compare two backups Which files and plugins changed between any two. 19 h ago1 h ago ### An assistant before a ticket Customers ask why their server lags first. AssistantActivity Assistant Ask about your server. It checks the server before answering.ChangesAsk first New conversation Why is TPS dropping in the evening? 2 h ago Ask about this server Performance, crashes, errors or settings. Why is my server lagging? What caused the last crash? Is anything misconfigured? Summarize today's console errors Ask about your server… Assistant use this period1,660,913 of 2,400,000 tokens ### Installed in one command It backs up the panel first and connects every server. AccountAccount settings ### Account settings Personal details, access, and preferences for your account. Profile Name and contact details Sign-in & security Password, MFA and sessions Appearance Theme, palette and density Notifications Email and message preferences Billing Plan, usage and payments Referrals Your link, signups and credit Organizations Your workspaces and access Connected panels Your own panels linked in owner mode Connected panels Run your own Pterodactyl panel? Connect it to respawn.sh and it connects your servers and installs the plugin for you. Install on your panel with one command It finds your panel, installs Blueprint if needed (it asks first) and connects the panel to this account. It backs up the panel files and database first and restores them if anything fails. Connect toSakura Network curl -fsSL https://api.respawn.sh/panel/install.sh | sudo bash -s -- --token 'rpit_8Kf2Qx7Lm4Vn'Copy Run it as root on the machine that hosts your panel. Keep it private: it works once, until 08:52 AM. Check the script before running it ### Every customer in one console Servers, usage and activity per customer. ### Customers The players and teams using respawn.sh through your panels. Search by name or panel IDUpdated just now CustomerServersAI tokens this monthLast active Sakura NetworkPanel ID panel-user-10041.4M tokens16 min ago Maple CraftPanel ID panel-user-1042610K tokens5 h ago Ember RoleplayPanel ID panel-user-1171280K tokens14 h ago Pixel IslesPanel ID panel-user-101195K tokens2 days ago ### Your brand, billed by usage White label on every tab and email. Usage flows into WHMCS or Paymenter, so you bill customers your way. ## Built for what agents can do next The pieces that let an AI operator act on a real server safely, and what lands next. Coming very soon ### Sandboxes for agents Before a change touches your server, Veyra will boot a private, isolated copy of your latest backup, apply the change and watch it, so you get a pass or fail before anything is real. Coming very soon ### Give your agents eyes Veyra joins the copy as a player, looks at what it built or changed, and checks it in game before it asks you to approve. Live ### Herd immunity When a plugin update starts crashing servers anywhere in the fleet, every other server hears about it first, with the fix that worked. Coming very soon ### Describe your server Say what you want in a sentence. Veyra plans the plugins, settings and staff ranks, and installs them once you approve, with a backup first. Live ### Playbooks you can share One-click automations built from checked actions, with approvals where they matter. Use your own or install one from the marketplace. Coming very soon ### Veyra in your Discord Status, player counts, incident alerts and approvals with a button, right where your staff already are. ## Questions Anything else? Talk to us (https://respawn.sh/contact) . What does the Free plan include? One server and 2 GB of encrypted snapshot storage, with daily to weekly snapshots kept for 7 days. Paid plans add servers, storage, more frequent snapshots and longer history. Where can I use Veyra? In your respawn.sh dashboard, for one server or the whole organization, and in the respawn.sh tab of hosting panels that offer it. Does Veyra act on its own? Only inside limits you set. By default it suggests, and owners choose what each member may ask it to do. Restarts, restores, console commands, file reverts, kicks and bans always wait for a person, whatever the settings say. What does Veyra remember, and can I remove it? Short facts about your organization, each server and you, saved from conversations and findings. You can see every one in the memory panel and remove any of them. Which games are supported? Minecraft is fully supported: Paper and Purpur 1.21.1 to 1.21.11, Folia 1.21.4 to 1.21.11, plus Leaves, Pufferfish, Canvas and Spigot, on Java 21 (Canvas needs Java 25). Hytale and FiveM servers get backups, restores, telemetry, crash signatures and Veyra; the profiler, player session analytics and automatic plugin updates are Minecraft only for now. How are my backups protected? Each snapshot is encrypted with AES-256-GCM on your server before it leaves, and sent over TLS. Snapshots are deduplicated, so only changed data is uploaded. How does a restore work? Pick a snapshot in the dashboard. On Minecraft with plugin 1.4.0 and a connected Pterodactyl or Pelican panel, Restore now takes a safety snapshot, stops the server cleanly, swaps the files, starts it again and checks it for 15 minutes. Otherwise the plugin stages it and applies it the next time the server starts, before worlds load. You can watch progress, cancel or resume. What does early access mean? The feature is built and running for a small test group, and is not yet open to everyone. I run a hosting company. Can my players use this? Yes. One command adds the respawn.sh tab to your Pterodactyl, Pelican or Calagopus panel. You pay respawn.sh by usage, and your players see their usage, never prices. ## Back up your first server today The Free plan covers one server with 2 GB of encrypted storage. No card needed. Back up your first server (https://auth.respawn.sh/signup) Ask a question (https://respawn.sh/contact)