Vape Filling Machine Comparison: Throughput, Automation, and Accuracy Across Leading Systems
If you are shopping for a vape cartridge filling machine right now, you are probably staring at a pile of spec sheets that all sound the same. Every vendor claims speed, precision, and easy cleaning. Almost none of them publish their numbers under the same conditions, which makes an honest side-by-side comparison surprisingly hard.
This guide fixes that. It puts the published specifications for the leading systems from Xylem Robotics, Thompson Duke Industrial, and Vape-Jet in one place, with sources, and then walks through how to test those claims against your own oil, hardware, and staffing before you sign a purchase order.
One caveat up front: this is a comparison guide, not a universal ranking. There is no single best vape cartridge filling machine for every producer, because the right choice depends on your concentrate, hardware, staffing model, facility, and downstream plans. What follows is a repeatable way to compare machines on the criteria that actually decide production outcomes.
What a Vape Filling Machine Does, and Why the Whole Workflow Matters
A vape filling machine is automated manufacturing equipment that dispenses cannabis oil into vape cartridges, pods, or all-in-one disposable devices at a target volume or weight. If you are new to the category, the Discover Vape Filling Machines for Cannabis Manufacturing guide provides an in-depth look at cannabis-specific filling considerations.
A meaningful purchase decision considers the entire workflow, not just fill speed. That includes cartridge loading, filling, capping, inspection, cleaning, packaging handoff, operator involvement, and format changeover. A machine that fills fast but bottlenecks at capping or cleaning does not raise your true output. It just moves the constraint somewhere you have not measured yet.
As you evaluate, keep three forms of evidence separate. Vendor-reported specifications tell you what a machine is designed to do. Live demonstrations show it running with materials in a controlled setting. Factory or site acceptance criteria bind the agreed results into your purchase contract. Do not treat a marketing claim as an independently verified finding, and do not let a demonstration stand in for a written acceptance test.
Published Specifications at a Glance
The table below compiles vendor-published specifications for six systems commonly on cannabis producers’ shortlists. Every number comes from the manufacturer’s own product page, press release, or an on-record trade interview, cited at the end of this article. None of these figures have been independently verified here, and every one should be reproduced with your own oil and hardware before purchase.
| System | Automation class | Published throughput | Published fill accuracy | Fill temperature range | Loading method |
|---|---|---|---|---|---|
| Xylem X4 | Fully automatic filler and capper | 1,500 to 1,800 units/hour with one operator; roughly 1,600/hour when flow-connected to a bagger [3][5] | Fill-weight deviation of 0.25% or less | 38°C to 95°C | No trays; carts and mouthpieces are bulk-loaded and machine-sorted |
| Xylem AX | Automatic filler (AI-enabled) | 500 to 700 units/hour | Fill precision of 1% | 40°C to 95°C | Machine vision, no specialized trays |
| Thompson Duke IZR | Automatic filler | 30+ fills per minute, roughly 1,800/hour [6] | Under 1 g product loss per 252-cartridge tray | Weight, temperature, and speed are controlled via HMI | Trays |
| Thompson Duke SVH | Semi-automatic, operator-stepped | Roughly 25 fills per minute with a skilled operator on a consistent formula; multi-SKU runs average lower [7] | Operator-stepped dispense control | Heated reservoir | Manual placement, operator steps for each fill |
| Vape-Jet 4.0 | Automatic filler | 1,200+ carts/hour with one operator; 5,000 to 10,000 devices per 8-hour shift [8] | ±0.25% repeatability; 0.001 mL dispense resolution with machine-vision QC | Three programmable heat zones; no numeric floor published — claims the lowest operating temperatures of any filling system | Trays |
| Vape-Jet Jet Fueler 3.0 | Semi-automatic | 4,000 to 8,000 carts per 8-hour shift, roughly 500 to 1,000/hour [9] | Dispense range 0.001 to 5.0 mL, third-party calibrated pump | Three independent PID heat zones | Manual placement, operator steps for each fill |
A few things jump out at you from this table once the numbers are side by side.
First, the fully automatic tier clusters at roughly 1,200-1,800 units per hour. The Xylem X4 stands alone due to its ability to fill and cap in-line. Thompson Duke’s IZR publishes a comparable hourly rate, with Vape-Jet 4.0 close behind, but both of those figures describe filling only. The X4 takes loose carts and mouthpieces with no trays at all, sorts parts internally, and caps inline in roughly three seconds per unit, which is why its throughput figure describes filled and capped output rather than filling alone.
Second, whether capping is included changes what the throughput number means; the X4 figure includes sorting, filling, and capping in a single process. The IZR and Vape-Jet 4.0 figures describe filling; capping runs on separate equipment (Thompson Duke sells pneumatic presses for that step, and Vape-Jet sells a separate capper). If your comparison spreadsheet has a single throughput column, you are comparing different amounts of work.
Third, low-temperature capability has become a real differentiator because of live resin and rosin. Xylem publishes numeric floors: 38°C on the X4 and 40°C on the AX. Vape-Jet does not publish a number; it claims the lowest operating temperatures of any filling system, supported by its nitrogen-pressurized fluid path. A numeric floor and an unquantified claim cannot be compared on paper. Thick cold rosin behaves nothing like warm distillate in a fluid path, so either statement only matters if the machine sustains rated speed at your temperature with your formulation. That is a demonstration item, not a spec-sheet item.
The Comparison Framework: What to Ask Every Vendor
Use this as your primary scannable framework. For every criterion, request each vendor’s documented specification, then require a demonstration using your actual oil and hardware. A specification you cannot reproduce with your own materials is a starting point for testing, not a result.
When you ask about throughput, require the vendor to state sustained units per hour, operator count, oil formulation, cartridge format, target fill volume, and whether that rate includes loading and capping. A number without those variables is not comparable to anything.
Market options span fully automated, automated, and semi-automatic systems for different operating needs, as the Cannabis & Tech Today roundup shows in its comparison [1]. Match the automation class to your production stage rather than assuming more automation is always better.
| Evaluation criterion | What to measure | What to ask every vendor to demonstrate | Why it affects the purchase decision |
|---|---|---|---|
| Throughput | Sustained units per hour with your oil, format, and fill volume | A timed run at the quoted rate, stating operator count and whether loading and capping are included | Peak rates rarely hold across a shift; sustained rate sets real output |
| Automation scope | Whether the system is a standalone filler, a filler-capper, or a broader line | The exact steps the machine performs versus the steps a worker performs | Determines labor, transfers, and where bottlenecks move |
| Fill accuracy and consistency | Fill-weight deviation across a realistic run | Test method, sample size, target fill weight, acceptance threshold, and full results | Deviation drives label-claim compliance and product waste |
| Integrated capping | Supported closure styles, capping time, torque, or press validation | Whether capping is inline or at a separate station, and cap-quality results | Capping is a common bottleneck and a leak-risk point |
| Footprint and modularity | Installed dimensions, utility needs, mobility, clearance | The space required for staging and downstream equipment, not just the cabinet | Undersized planning stalls installation and flow |
| Changeover | Time to switch between your actual formats | A live change including fixtures or trays, programming, and first-pass validation | Multi-SKU operations lose hours to slow changeover |
| Downstream integration | Compatibility with cleaning, QC, inspection, packaging, and production data | Whether each handoff requires a worker | Manual handoffs cap the throughput of the whole line |
| Cleaning and QC | Cleaning procedure, access, and inline quality checks | The cleaning steps between runs and what QC occurs before packaging | Cleaning time and rework reduce effective daily output |
| Commissioning and training | Installation scope, operator training, and manuals | A documented install plan and training curriculum | Poor commissioning delays first production and raises error rates |
| Ongoing service and support | Response commitments, spare-parts availability, remote support, preventive maintenance | Written response times and a maintenance schedule | Downtime cost depends on how fast support and parts arrive |
How the Three Manufacturers Actually Differ
Thompson Duke Industrial: tray-based filling in OEM packaging
Portland-based Thompson Duke has been in cannabis filling longer than almost anyone. Its legacy MCF1 semi-automatic filler sold over 1,200 units worldwide and filled well over half a billion cartridges before the company sunset it in July 2024 in favor of the SVH [7]. The current lineup pairs the fully automatic IZR with the semi-automatic SVH, both cETLus-certified for North America, with CE variants for the European market.
The IZR’s defining design choice is to fill cartridges directly into the manufacturer’s packaging trays, up to 252 devices per tray, with a published product loss of under 1 gram per tray and a maximum cartridge height of 5.25 inches [6]. The operator loads a tray and collects a completed tray. Power requirements are modest: 120VAC on a dedicated 15A circuit, drawing 250W. Capping is a separate step handled by Thompson Duke’s pneumatic presses, which is the main workflow difference versus an inline filler-capper.
The SVH is the multi-SKU workhorse of the line. Thompson Duke publishes roughly 25 fills per minute with a skilled operator running a consistent formula, and is candid that multi-SKU production averages lower because of changeovers [7]. It also accepts the Little JIM joint-infusion bolt-on, which converts the same bench into a pre-roll infusion station, a genuinely useful trick for producers running both product categories on one budget.
Vape-Jet: nitrogen-assisted precision with machine-vision QC
Vape-Jet’s flagship VJ4.0 produces 1,200+ carts per hour with a single operator, and the company publishes a shift range of 5,000 to 10,000 devices per 8-hour shift [8]. Its signature engineering choices are a nitrogen-pressurized fluid path to suppress oxidation, three programmable heat zones with a stated claim of the lowest operating temperatures of any filling system (no numeric floor is published), and AI-guided machine vision that aligns the dispense tip with each cart and performs a QC scan on every unit. The company publishes a dispense resolution of 0.001 mL and states that one operator can run up to four machines simultaneously.
The semi-automatic Jet Fueler 3.0 publishes 4,000 to 8,000 carts per shift, with a third-party-calibrated pump dispensing 0.001 to 5.0 mL and a clean-in-place cycle that completes in under 10 minutes using about 200 mL of alcohol [9]. Its fluid path uses fluoropolymers, Viton, borosilicate glass, stainless steel, and ceramic.
Vape-Jet loads hardware from manufacturer trays of 50 to 100 units placed directly on the filling platform. Like the IZR, it is a filler; capping is a separate station in the Vape-Jet ecosystem, so quoted throughput describes filled, not filled-and-capped, output.
Xylem Robotics: flow manufacturing with inline capping
Houston-based Xylem Robotics took a different path to market. The company spent roughly five years building its machines for internal use in its own California Type 7 volatile-solvent facility before selling them commercially, which means the current X4 is a fourth-generation design shaped by production floors rather than trade-show demos [3][4].
The X4 is a flow manufacturing system that sorts, fills, and caps in a single sequence, operated by a single operator. It produces 1,500 to 1,800 filled and capped units per hour, or roughly 1,600 per hour when flow-connected to a bagger, for end-to-end output, with each cart exiting filled, capped, cleaned, and bagged [3][5]. There are no trays anywhere in the process; operators bulk-load loose cartridges and mouthpieces, and the machine handles orientation. Capping takes roughly three seconds per unit across press, click, and screw closures, and the machine checks every cartridge center post; combined with instant capping, this is how Xylem addresses leak risk. The published fill-weight deviation is 0.25% or less; the fill-temperature floor is 38°C; and the sealed resin path runs stainless-steel tanks and lines with ceramic pumps, with an optional nitrogen or argon overlay for oxidation-sensitive resins. Many times the filled carts are then passed into automated packaging machines directly saving more labor and time in pack-off processes.
The AX is Xylem’s foundational filler for small and medium operations. It publishes 500 to 700 units per hour, 1% fill precision, a 40°C to 95°C operating range, and machine-vision hardware that handles cartridges, pods, and all-in-one devices from any manufacturer, up to 4.7 mL, with form-factor swaps in minutes and no specialized trays.
Throughput, Accuracy, and Temperature: How to Validate the Claims
Compare throughput claims only when the variables match. Material type, viscosity, target volume, cartridge format, staffing, and whether capping is included must be equivalent across quotes. A rate for warm distillate in a simple 510 cartridge will not predict the output for cold rosin in a wide-body device. Xylem’s own product page illustrates the spread: most X4 run speeds average around 1,650 units per hour, with some premium live resin formulations reaching 1,850, meaning formulation alone can move the number by more than 10% on the same machine [5].
Measure sustained throughput, not peak output
A peak rate captured for a few minutes rarely holds for the duration of a shift once loading, cleaning, and interventions are accounted for. Ask for a sustained rate over a defined run and record every stoppage. A useful sanity check: divide the vendor’s per-shift figure by eight and compare it to the per-hour headline. Vape-Jet’s published shift range of 5,000 to 10,000 works out to roughly 625 to 1,250 per hour, which brackets its 1,200+ hourly headline only at the top of the range — a useful reminder that shift averages already include the stoppages a peak rate ignores. Where a vendor’s shift math and hourly math diverge further than that, ask why.
Validate fill consistency with an acceptance test
Fill consistency is the criterion most closely tied to compliance and waste, and it is also the area where published figures are least comparable. Xylem publishes a fill-weight deviation of 0.25% or less on the X4. Vape-Jet publishes a ±0.25% repeatability rate alongside a 0.001 mL dispense resolution. Thompson Duke publishes product loss per tray (under 1 g per 252 carts on the IZR). Even where the percentages look identical, deviation and repeatability can be measured against different baselines, sample sizes, and test conditions, and none of the vendors publish their test methods. Define your own target fill weight, sample size, and acceptance threshold, then measure the actual gravimetric deviation across a realistic production run on every machine you shortlist, so the comparison is finally apples-to-apples.
Match the temperature capability to the concentrate
Temperature capability decides whether a machine can handle your concentrate without degrading it. The X4 publishes a 38°C floor and the AX publishes 40°C, both in the range where terpene preservation in live resin and rosin becomes practical. Vape-Jet publishes no numeric floor, claiming instead the lowest operating temperatures of any filling system, so that claim has to be quantified in a demonstration rather than compared on paper. The question your demonstration must answer is not whether the machine reaches that temperature, but whether it sustains its rated throughput and accuracy at that temperature with your specific formulation, because viscosity climbs steeply as temperature falls.
Run a written acceptance test
Build the test on this checklist: an agreed throughput window, a defined run duration, a named operator count, your actual hardware and oil, a fill-weight sampling plan, rejection criteria, leak checks, cap-quality checks, and documentation of any downtime or interventions. Then convert the agreed outcome into a purchase acceptance criterion in the contract. Xylem states it takes new machines onsite to clients before release because lab conditions differ from production [3]; hold every vendor, Xylem included, to that same standard on your floor.
Automation, Capping, and Downstream Integration
Automation comes in distinct configurations, and no single configuration fits every facility. A standalone filler dispenses oil and hands off for capping, which suits operations that already own capping capacity. A sequential fill-and-cap system does both in order. A flow-connected system links to cleaning and packaging, so the product moves through with fewer touchpoints.
The practical difference shows up in headcount and handoffs. On a filler-plus-separate-capper line, filled carts wait between stations, and a worker moves them. On the X4’s inline configuration, a cart is capped roughly three seconds after filling and can pass directly to downstream packaging without further worker action. Every eliminated handoff removes a place where the line rate can drop and where a cart can leak, become contaminated, or be miscounted.
When you talk to a vendor about capping, ask four things. Does the quoted throughput include capping or only filling? Do caps load automatically or by hand? What QC happens before packaging? Do filled, capped units transfer to packaging without a worker touching them? A short checklist of these questions with each vendor beats trusting a single headline rate.
Footprint, Utilities, and Changeover
Evaluate the installed footprint, not just the cabinet dimensions. Account for floor space, utility access, material staging, cleaning access, packaging adjacency, and safe operator movement.
The published requirements are concrete enough to plan against. The Xylem X4 lists system dimensions of 71 × 43 × 31 inches, a recommended working space of 10 × 10 feet, 120V/11A power, and 70 to 90 psi compressed air at 4.6 CFM. The Xylem AX lists 36 × 32.5 × 30 inches, a 5 × 5-foot working space, 120V/10A power, and optional 70 psi air at 1.6 CFM. The Thompson Duke IZR runs on a standard 120VAC dedicated 15A circuit at 250W [6]. None of these machines demands three-phase power or unusual utilities, which keeps facility prep manageable. Still, compressed-air capacity is the item most often missed in planning for the X4 class.
Changeover deserves its own timed test. Tray-based systems trade fixture cost and tray-swap time for repeatable positioning; vision-based loose-part systems trade that for faster format swaps. The AX publishes form-factor changes within minutes without specialized trays, and Thompson Duke positions the SVH specifically for producers running three or more SKUs in a single setup because there is less hardware to clean between runs [7]. Whichever architecture you choose, time a full format change between your actual SKUs, including programming and first-pass validation, during the demonstration.
Prove Hardware Compatibility Before You Commit
Filling equipment compatibility must be validated with the precise cartridge hardware, fill volume, temperature, and closure method before you commit a production order, as detailed in the Finished Goods compatibility explainer for piston, pump, and gravity systems [2]. Test actual production samples under real operating conditions and obtain written confirmation of compatibility from your hardware supplier before volume purchasing.
The published compatibility envelopes differ meaningfully. The Xylem AX accepts cartridges, pods, and all-in-one devices from any manufacturer up to 4.7 mL. The X4 covers all decarbed resins, including live resin and distillate, and all 510-thread top-filled cartridges with press, screw, or click closures. The Thompson Duke IZR accepts hardware up to 5.25 inches tall in standard OEM trays of up to 252 devices [6]. Vape-Jet publishes compatibility information for 510-style devices, pods, disposables, capsules, and syringes [8]. Work through the practical checklist item by item during evaluation: fill-path clearance, fill volume, required temperature, repeatable fixture holding, capping pressure, mouthpiece installation, special handling for postless or wide-body devices, settling time, and documented line-speed limits. Each item can fail independently.
Which Xylem System Fits Your Production Stage?
Position these systems based on documented workflow fit rather than declaring either one universally superior.
| System | Published workflow fit | Published throughput | Hardware and concentrate compatibility | Automation and integration |
|---|---|---|---|---|
| Xylem AX | Foundational cart-filling for small and medium operations | 500 to 700 units per hour | Cartridges, pods, and all-in-one devices from any manufacturer up to 4.7 mL | Machine vision, no specialized trays; form-factor swaps within minutes; fill precision 1%; 40°C to 95°C |
| Xylem X4 | High-speed flow manufacturing, filling, and capping in sequence | 1,500 to 1,800 units per hour | All decarbed resins, including live resin and distillate; 510-thread top-filled carts with press, screw, or click closures | Standalone or flow-connected with cleaning and packaging; ~3-second capping; fill-weight deviation ≤0.25%; 38°C to 95°C |
The AX suits producers who need reliable, precise filling in a constrained footprint with flexible hardware handling. Its tray-free machine-vision design lets a small team swap formats quickly, and its 500- to 700-units-per-hour range fits pilot and growing production. The GMP-oriented pressurized stainless-steel path supports quality requirements without the space and integration demands of a full line.
The X4 suits producers who need high volume without adding headcount and who plan to connect cleaning and packaging. It sorts, fills, and caps in one process with one operator, translating to roughly 10,000 units in an eight-hour shift [4], and its low fill temperature and sealed resin path support live resin at scale. Validate its published throughput and accuracy with your own materials before committing.
A Decision Path for Small, Scaling, and High-Volume Manufacturers
Base your path on operational needs, not on unsupported price bands or competitor rankings.
Small or pilot-stage operations should prioritize a constrained footprint, flexible hardware handling, manageable operator training, output that matches real demand, and a documented path for later expansion. In this tier, the realistic candidates are the semi-automatic machines (Thompson Duke SVH, Vape-Jet Jet Fueler) and the compact automatics (Xylem AX). A machine that fits your current space and staffing while leaving room to grow avoids an early replacement.
Scaling multi-SKU operations should prioritize repeatable changeover, validated hardware compatibility, fill consistency, reduced staffing, and the ability to connect future downstream processes. This is where architecture choices bite: tray-free vision loading (AX, X4), OEM-tray filling (IZR), and manufacturer-tray platform loading (Vape-Jet) each handle format churn differently. Time the changeover on your own SKUs.
High-volume and integrated production environments should prioritize sustained, validated throughput; inline capping; reduced touchpoints; integration with cleaning, QC, and packaging; uptime planning; and support coverage. The X4’s published 1,500 to 1,800 units-per-hour range, its sequential filling and capping, and its flow-connected configuration, reported in the CEN at MJBizCon coverage and on the X4 product page, illustrate the requirements a high-volume facility should validate against its own run [3][5]. Xylem’s Houston-based automation work reflects that high-volume focus [4].
To move forward, provide your cartridge hardware and oil profile to each shortlisted vendor for a demonstration and a documented production evaluation.
Citations
[1] https://cannatechtoday.com/best-vape-cart-filling-machines
[2] https://www.finishedgoods.com/blog/cannabis-hardware-filling-system-compatibility-thompson-duke-dds-atg-and-vape-jet
[3] https://www.cannabisequipmentnews.com/video/video/22931898/cen-at-mjbizcon-xylems-x4-filling-machine-makes-up-to-1800-carts-per-hour
[4] https://cannabisnow.com/xylem-robotics-automated-innovations-improve-cannabis-tech
[5] https://xylemtech.com/automated-cart-filling-machine/
[6] https://thompsondukeindustrial.com/machines/izr/
[7] https://thompsondukeindustrial.com/mcf1-semiautomatic-vaporizer-filling-machine/ and https://thompsondukeindustrial.com/news-release-out-with-the-old-in-with-the-new-superior-filling-with-the-svh-and-the-sunset-of-the-renowned-mcf1/
[8] https://vape-jet.com/vape-jet/
[9] https://www.usalab.com/vape-jet-jet-fueler-3-0-semi-automatic-vape-cartridge-filling-machine/
Frequently Asked Questions
Published rates for fully automatic systems cluster between roughly 1,200 and 1,800 units per hour. The Xylem X4 publishes 1,500 to 1,800 filled and capped units per hour with one operator, while the Thompson Duke IZR and Vape-Jet 4.0 publish comparable rates for filling only, with capping handled on separate equipment. Because the machines measure different amounts of work, the honest answer requires a sustained timed run with your own oil and hardware rather than a headline comparison.
Not always, and it changes what the number means. The X4’s figure covers sorting, filling, and capping in one inline sequence. The IZR and Vape-Jet figures describe filled output; capping runs at a separate station. When you compare quotes, require every vendor to state whether the rate covers filled units, filled-and-capped units, or packaged goods.
Lower is generally better for terpene preservation. Among published numeric floors, the X4 lists 38°C and the AX lists 40°C; Vape-Jet claims the lowest operating temperatures of any filling system without publishing a number. Because viscosity climbs steeply as temperature falls, the deciding test is whether a machine holds its rated speed and accuracy at your target temperature with your specific formulation.
Run a timed demonstration with your actual oil, hardware, and target fill volume, record the operator count and every stoppage, and count accepted finished units rather than machine cycles. Then write the agreed sustained rate, fill-weight sampling plan, and rejection criteria into the purchase contract as acceptance criteria.
Only to the extent it mirrors your conditions. It must use your oil, hardware, fill settings, and operator count, and run long enough to include cleaning and changeover. A short run with vendor-supplied materials shows capability, not your sustained daily output.
Confirm written service response times, remote support options, and spare-parts stock before purchase, and keep critical wear parts on hand to limit downtime. For turnover, keep manuals and a documented training curriculum on-site so a new operator can be brought up to speed without relying on one person’s knowledge.