Klipper tunes individual printers brilliantly, but it gives you nothing for running several of them as a unit. Fleet management is a layer you add on top of Klipper — not a firmware feature. That single fact explains most of the confusion around Klipper fleet management: there is no built-in “farm mode”, so everything you want — one dashboard, a shared queue, per-part costing, failure alerts — comes from tooling you bolt onto Moonraker’s API. Get the architecture right first and the rest follows.
Every Klipper machine runs the same three-part stack: the Klipper firmware (a host process plus the printer’s MCU), Moonraker (a Python 3 web server exposing a RESTful API and websocket), and a web UI such as Mainsail, Fluidd or KlipperScreen. Moonraker is the bit that matters for fleets. It bridges Klipper to anything external over a Unix domain socket to Klippy and a TCP port to the network. Every fleet tool below talks to that API. If a printer doesn’t expose Moonraker — stock Marlin, for example — it can’t join a Klipper-style fleet without a translation layer.
What Klipper gives you, and where fleet tooling starts
By default Moonraker listens on TCP 7125 and binds to 0.0.0.0 so it accepts connections from any interface on your LAN. That’s the hook every dashboard and dispatcher uses. In a mixed fleet you’ll be juggling protocols: Klipper machines on Moonraker over 7125, Bambu Lab printers over MQTT (ports 1883 and 8883), PrusaLink and OctoPrint over their own REST APIs. Any tool claiming to run a mixed fleet has to speak all of those — and most self-hosted options only speak Moonraker.
The useful way to think about a Klipper farm is as three separable jobs, because the tool that nails one rarely does the other two:
- Monitoring — live status, temperatures and cameras for every machine in one view.
- Dispatch — a queue that holds jobs and decides which printer gets the next one.
- Production — parts libraries, works orders, stock, per-part cost, despatch paperwork.
Native Klipper tooling handles the first job partially and the other two not at all. Knowing which layer you actually need stops you installing five things that overlap.
The native options: Mainsail and Fluidd multi-printer
Both Mainsail and Fluidd list multi-printer support, and for a handful of all-Klipper machines that may be all you need. But be clear about what “multi-printer” means here: it’s a switcher, not an aggregated dashboard. Fluidd connects to one printer at a time and fully reinitialises when you switch between instances in the browser tab. Mainsail leads its feature list with multi-printer management and gives you a tidier switch, but it’s still fundamentally per-printer control from one UI.
The hard limitation is that neither gives you a shared job queue. There’s nothing that hands the next file to whichever printer is free. You route every job yourself. For two or three printers that’s fine. For six machines running the same part, it’s six manual starts — and that’s where people start looking for a dispatcher.
Running several Klipper instances on one host
You don’t need a Raspberry Pi per printer, though that’s the classic pattern. A single more powerful host can run multiple Klipper and Moonraker instances — one pair per printer — each on its own port, with its own config and logs.
Do not try to set this up by hand. The honest answer is that without solid Klipper and Linux experience, manual multi-instance setup is painful. Use KIAUH (Klipper Installation And Update Helper), which does the legwork: it installs and configures multiple Klipper and Moonraker instances on one system, each isolated with its own service file, config files, port allocation and log file.
A few behaviours worth knowing before you commit:
- Each printer instance gets its own folder in your config location, and each service name changes slightly in a multi-instance install (single-instance KIAUH setups keep their original names).
- Each instance produces its own klippy.log, accessible through its Moonraker page like any single printer — as long as you’ve set the paths correctly.
- You can add or remove instances later; it’s no longer a one-shot decision.
This gets you several printers off one box with per-printer UIs. It still doesn’t give you a queue or a farm view — for that you add a dashboard on top.
Self-hosted fleet dashboards and dispatchers
Here’s where the advice online goes stale, so a correction: OctoFarm is effectively dead for Klipper. Its last stable release is v1.7.3 from October 2022, with a v1.8.0-beta in December 2022 and nothing since — and it only ever managed OctoPrint instances. A Klipper farm gets nothing from it today. If a guide recommends OctoFarm for Klipper, it’s out of date. Our own OctoFarm alternative comparison walks through what has replaced it.
The live self-hosted options split by what layer they cover:
| Tool | Printers supported | Queue / dispatch | Licence / hosting |
|---|---|---|---|
| Obico | OctoPrint, Klipper/Moonraker | No queue — monitoring and AI failure detection only | AGPL-3.0; obico.io or self-host via Docker |
| FDM Monster | OctoPrint, Klipper via Moonraker, PrusaLink, Bambu LAN | Brand-neutral fleet layer | Open source, self-hosted |
| print-farm-manager | Moonraker, OctoPrint, PrusaLink, Bambu, Elegoo | G-code dispatch to next idle printer (no model library) | MIT, free |
| LunarPrintfarm | Klipper/Moonraker, OctoPrint | Queueing and distribution | Open source |
Obico is a solid monitoring and remote-access layer with AI failure detection, but it has no print queue and no continuous-printing automation — you start and route each job yourself. print-farm-manager does dispatch: projects hold parts and uploaded G-code, and its engine routes each job to the next suitable idle printer. One caveat to plan around — every queued job targets a single printer, so running one part across six machines means six queued jobs, not one broadcast. For a mixed fleet you’ll want a brand-neutral layer such as FDM Monster, Obico or a hosted platform rather than relying on Mainsail’s Klipper-only view.
Remote access across the fleet
OctoEverywhere gives free, private, unlimited remote access to your full Mainsail, Fluidd, OctoPrint and Bambu web portals from anywhere, plus AI print-failure detection. The operational gotcha for fleets: it’s installed per instance, so a firmware or OS update can break the connection until you re-run the install script. One user running two Pis updated both, found neither printer would connect through OctoEverywhere, and had to SSH into each and re-run the installer. Budget for that maintenance if you go per-instance.
Commercial and hosted platforms
If you’d rather not self-host the queue, the hosted options are mature. SimplyPrint works with Bambu Lab, Prusa, Creality, Klipper/Moonraker/Mainsail/Fluidd, OctoPrint and direct USB, and adds a print queue, cloud slicer, filament inventory, teams and SSO. Its AutoPrint (continuous printing) automatically starts the next job when one finishes, clearing the bed however suits each printer — toolhead push-off, plate-changer mods like SwapMod and JobOx, belt printers, robot arms, the API, or a manual mode where you clear by hand and mark the bed ready. Its AI bed check watches any camera-equipped printer, catches a failing print, and confirms the bed is clear before the next job starts. If you’re weighing it up, our SimplyPrint alternative comparison lays out the trade-offs. 3DPrinterOS supports 200+ models including Klipper machines, with queue management, cloud slicing and real-time monitoring.
Where Printer Farm fits — running the whole job end to end
Monitoring and dispatch are the first two of the three jobs. If you’re selling printed parts or running a production workflow, you need the third — and that’s where Printer Farm is built differently. It runs the whole pipeline end to end rather than just watching printers.
A web order lands — from WooCommerce (one click, read-only key), Shopify, or your own custom website connected through the farm’s shop API (a token, no plugin) — and is pulled in every 10 minutes and matched to part numbers by SKU. Stock is used first; one click raises a works order for exactly the shortfall, carrying the due date and customer reference. Slicing runs through the real OrcaSlicer engine — in the browser Cloud Slicer with full Printer, Filament and Process presets, imported Orca profiles and STEP conversion, or plates arrive from ATN Slicer as revisions. The queue dispatches each job to the first free compatible printer with the right filament loaded — the shared-queue behaviour native Klipper tooling can’t do. Cameras with AI failure detection watch each print and alert you. Good prints bank into stock and onto the works order, spools are deducted gram-by-gram as prints run, and a true cost per part is recorded (mixed plates split by weight). At despatch you confirm the bed is clear, press Despatch, and a numbered delivery note is raised — and for a custom website connected through the shop API, the order is marked shipped and the customer emailed automatically.
On the Klipper front it works with Creality and any Klipper/Moonraker printer, plus Bambu Lab in LAN mode, FlashForge, OctoPrint, PrusaLink and Duet — so your mixed fleet runs on one dashboard. A small agent runs on a Windows or Linux computer that stays on and bridges the printers; your files stay on your own machine, and slicing runs on your PC through ATN Slicer (Workshop and Production can also slice on our servers). If this is your situation, our guide to starting a 3D print farm and the piece on running one queue across a fleet go deeper.
FAQ
Does Klipper have built-in fleet management?
No. Klipper optimises individual printers and exposes an API through Moonraker (default TCP 7125), but it has no native system for managing printers in parallel. Fleet features — a shared dashboard, a queue, costing — all come from software layered on top of Moonraker.
Can Mainsail or Fluidd manage multiple printers?
They can switch between several Klipper instances from one interface, but it’s a switcher rather than an aggregated dashboard — Fluidd fully reinitialises on each switch — and neither offers a shared job queue that hands the next file to a free printer. For a few all-Klipper machines that’s often enough; for dispatch you need a dedicated tool.
How do I run several Klipper printers on one Raspberry Pi?
Use KIAUH to install multiple Klipper and Moonraker instances on one host. Each gets its own service file, config, port and klippy.log, and you can add or remove instances later. It’s difficult to do by hand, which is exactly why KIAUH exists.
What’s the best tool for a mixed Klipper and Bambu fleet?
You need a brand-neutral layer, because most self-hosted dashboards only speak Moonraker. FDM Monster, Obico, SimplyPrint and Printer Farm all handle mixed fleets including Bambu in LAN mode — Printer Farm adds the order-to-despatch pipeline, works orders, stock and true per-part costing on top.
Ready to run one or a whole mixed fleet as a production line? Start the open beta with a free 14-day Workshop trial — no card required. Still deciding between tools? The fleet software buyer’s guide covers how to choose.
Related: 3D Print Defect Identifier: Read the Symptom, Name the Fault, Fix the Cause
