A soft robotic gripper holds parts with flexible silicone or elastomer fingers that bend around the object instead of clamping it. It's the right tool for irregular, delicate or wet products like raw protein, produce and baked goods. It's the wrong tool when you need tight placement, high grip force or heavy payloads.
Buy a soft gripper when the product varies piece to piece and bruises, tears or crushes under a hard jaw. That's most fresh food, a lot of e-commerce picking and some thin-wall plastics. Don't buy one to load a fixture, press a fit or move a 5 kg casting. For those, custom gripper fingers on an electric or pneumatic jaw will hold position and force far better.
A soft robotic gripper is an end effector whose fingers are made of a compliant material, usually silicone or polyurethane, so they deform to match the part. Most are driven by air pressure, vacuum or a small electric motor, and they hold the part with friction spread over a wide contact patch.
The classic design is a pneumatic network finger. It's a row of hollow silicone chambers molded onto a thin layer that won't stretch. Pump air in and the outer wall of each chamber balloons while the inner layer stays the same length, so the finger has nowhere to go but to curl inward. Vent it and the silicone springs back open. That's the principle behind the mGrip fingers, and you can see it in the drawing below.
There are other ways to get the same result. Piab's piSOFTGRIP molds the fingers and a sealed vacuum cavity as one silicone piece, so applying vacuum closes the fingers and the vacuum level sets how hard they pinch. OnRobot's Soft Gripper skips air entirely and uses a built-in electric motor to push a spindle that flexes the silicone cups. Festo's DHAS fingers use the Fin Ray structure, a flexible polyurethane frame that wraps toward the part when you push on it, mounted on an ordinary parallel or angle gripper.
They all trade the same things. You lose a precise, repeatable finger position and a high clamp force. You gain a gripper that doesn't care much whether this chicken breast is thicker than the last one, and that spreads the load so it doesn't leave marks. If you're still choosing between gripper families, our guide on how to choose a robot gripper walks through the full decision.
The four soft gripper architectures you'll meet in quotes, with a published example of each.
Soft grippers work best on products that change shape or size from piece to piece and can't take a point load: raw meat and poultry, bakery items, fruit and vegetables, bagged goods and thin-wall plastic. They also suit high-mix picking where you'd otherwise change fingers for every SKU.
Fillets are wet, slippery and never the same shape twice. A vacuum cup leaks on the uneven surface, and a rigid jaw tears the meat. Multi-finger soft tools wrap the piece and hold it by friction over a big area. Washdown rating matters more than grip force here.
Rolls are light but crush easily, and a porous crust won't seal for vacuum. Two or four soft fingers lift from the sides with low contact pressure. Watch for flour and crumbs building up between fingers, which changes friction over a shift.
Diameter varies a lot within one lot, and bruises show up days later at the store. Soft fingers absorb the size spread without a program change. Check the grip range against your smallest and largest piece, not the average.
Poly bags of loose product slump into a new shape every time they're touched. An adaptive gripper with soft fingers pinches at the fullest point and doesn't need to know the exact outline. Pair it with a camera that finds a graspable spot.
Cups, vials and cosmetic tubes dent or crack under a hard jaw set high enough to hold them at speed. Soft fingers grip at a stiff band like the rim and spread the load. This is one place where a soft tool also beats vacuum on curved surfaces.
The same compliance that protects a tomato lets a pin tilt in the grip. Pressing a pin into a reamed bore, loading a lathe chuck or seating a bearing needs the part held square and repeatable. Use machined fingers with a V or pocket for that job.
Notice what the good fits have in common. The robot is mostly picking and placing into a tray, a pack or onto a conveyor, where a few millimeters of placement error don't matter. For harder picks out of a jumbled tote, the gripper is only half the cell, and the camera that finds the part is the other half. Our vision-guided robotics guide covers that side, and bin picking cells shows how we put the two together.
Most single soft grippers are rated from a few hundred grams to about 2 kg, and the rating depends on the part's shape and surface, not just its weight. Large multi-finger tools go further; Schmalz lists mGrip configurations up to 10 kg and workpieces up to 300 mm wide.
The chart is the most useful thing a soft gripper datasheet can show you, and most don't show it. OnRobot does. Same base, same silicone tool, and the rated payload falls from 2.2 kg on a 65 mm cylinder to 0.5 kg on a 65 mm ball. That's a 4.4× spread with nothing changed but the shape. The softer SG-a-S tool drops the cylinder rating to 1.5 kg, and the larger SG-b-H tool, which reaches a 118 mm grip, is rated 1.1 kg.
So when a sales sheet says "up to 2 kg," read it as "up to 2 kg on the friendliest shape we could find." Then ask what it does on yours. Wet product, oil and flour all cut the friction coefficient, and acceleration adds to the load. A robot that lifts at 1 g of extra acceleration is asking the fingers to hold twice the part's weight for that moment.
The upside is that soft fingers spread whatever force they apply. That's why they can hold a bread roll without leaving a dent, and the arithmetic below shows the size of the effect.
On a lot of food lines the first soft gripper trial goes great on Monday with fresh, cold, dry product and falls apart on Thursday when the product is warmer, wetter or coated. Run your trial across the real range: start and end of a shift, first and last of a batch, and after a sanitation cycle when the fingers are still damp.
μ and contact areas are assumptions for illustration; measure friction on your product. The 32 grips per minute figure is OnRobot's published maximum for the SG-a-H tool. For the general friction-grip formula and how vendors apply safety factors, see gripper finger design.
Soft grippers give up precision, grip force and wear life in exchange for gentleness. The part can shift or tilt in the fingers, they hold far less than a rigid jaw of the same size, silicone tears and abrades over time, and multi-finger tools take real effort to clean.
No soft gripper vendor we've found publishes a fingertip repeatability, and that silence tells you something. The motor or valve behind the fingers may be repeatable, but the part's position in a compliant grip depends on its size, its stiffness and how it was sitting when the fingers closed. Fine for dropping a fillet in a tray. Not fine for loading a CNC vise.
The squeeze is limited by how much the silicone can push before it just deforms more. That's why soft tools are rated in grams to low kilograms, while an electric parallel gripper the same size might be rated for several times that with a form-fit finger. If you're weighing actuator types more broadly, the trade-offs are laid out in electric vs pneumatic grippers.
Silicone is a consumable. OnRobot rates its silicone tools for at least 2 million cycles and its base for 3 years or 3 million cycles, which is a useful benchmark. At 30 picks a minute over two eight-hour shifts, 2 million cycles is about 70 production days, or a bit over three months of five-day weeks. Budget for spare fingers and a quick swap, and check whether they're sold individually.
Every finger, fitting and seam is a surface someone has to clean. Pick rate is the other quiet limit. Fingers take time to inflate and vent, and soft parts swing a bit after a fast move. A vacuum cup on a flat, sealed product will usually be faster and simpler, which is why we'd still reach for a vacuum gripper whenever the surface allows it.
For most buyers the short list is Schmalz mGrip for multi-finger food tools, OnRobot's Soft Gripper for an all-electric cobot option, Piab piSOFTGRIP for small vacuum-driven picks, and Festo's DHAS Fin Ray fingers when you want to add soft fingers to a gripper you already have.
| Product | Drive | Published capacity | Food-contact claim (as published) | Best for |
|---|---|---|---|---|
| Schmalz mGrip (ex Soft Robotics) | Pneumatic, or electric control unit | Up to 10 kg, workpieces 20 to 300 mm, up to 120 picks/min | "FDA-compliant" POM, silicone and stainless steel; IP69K | Protein, bakery, high-rate food lines |
| OnRobot Soft Gripper (SG) | Electric, no air supply | 0.5 to 2.2 kg by shape (SG-a-H); grip range 11 to 118 mm across tools | FDA 21 CFR 177.2600 and EC 1935/2004, tested for non-fatty foods | Cobot cells with no compressed air |
| Piab piSOFTGRIP | Vacuum | Objects up to Ø30 mm (30-3) or Ø40 mm (40-3, new in 2026) | Detectable silicone, FDA 21 CFR and EU 1935/2004 | Small confectionery and delicate items |
| Festo DHAS fingers | Passive, on a host gripper | Sizes 60, 80 and 120; retention force by curve | Refers to supplementary material information | Retrofitting an existing jaw |
| Festo DHEF | Pneumatic silicone cap | Objects Ø12 to 38 mm (DHEF-20) | Refers to supplementary material information | Odd shapes picked from above |
Table scrolls horizontally on small screens →
If you searched for "Soft Robotics mGrip," you've probably noticed the company's site has changed. In August 2024 J. Schmalz GmbH, the German vacuum specialist, bought the flexible gripping product assets from Soft Robotics Inc., including all soft finger gripper configurations, patents and know-how. Soft Robotics renamed itself Oxipital AI the same week and now focuses on AI vision for picking and inspection. So mGrip support, spares and new orders run through Schmalz, which also added a lower-cost mGrip Basic series.
That's good news for most buyers, honestly. Schmalz is a large, established supplier with a US subsidiary, and spare-parts continuity is the first thing we'd worry about with any niche gripper.
Don't compare the headline payloads; they're measured on different shapes and in different ways. Compare on your product. Ask each vendor, or your integrator, for a trial on 50 to 100 real pieces covering your full size range, and log drops, marks and cycle time. Ask what a replacement finger costs and how long a swap takes. Ask whether the food claim covers your product, fatty or not.
And check how the gripper talks to your robot. Several of these ship with cobot plug-ins, but a pneumatic tool on an industrial arm still needs valves, a control module and someone who knows where the air goes. If you're also looking at mainstream two-finger options, our Robotiq alternatives page and the Robotiq vs OnRobot comparison cover the wider cobot gripper market.
Some are approved for direct food contact, but only under specific regulations and for specific food types. In the US the usual reference for silicone is FDA 21 CFR 177.2600; in Europe it's the framework Regulation (EC) No 1935/2004. Hygienic design and cleaning are separate questions from material approval.
21 CFR 177.2600 covers rubber articles intended for repeated use, and it sets extraction limits with two solvents. Distilled water stands in for aqueous foods, and n-hexane for fatty foods. The fatty-food test is much harsher, which is why OnRobot's datasheet notes its silicone was tested and approved for non-fatty food objects. If you're handling bacon, cheese or anything oily, that footnote matters.
In the EU, Regulation (EC) No 1935/2004 is the framework for every food-contact material. It says materials mustn't transfer constituents in amounts that endanger health or change the food unacceptably. Plastics have their own detailed measure, Regulation (EU) No 10/2011. Silicone and rubber don't, so the European Commission notes national rules apply. That's why a vendor's EU claim usually cites 1935/2004 plus a test report, and you should ask for that declaration of compliance.
Piab describes the piSOFTGRIP material as detectable silicone. The point is that a torn fingertip fragment can be caught downstream instead of shipping in a package. If your plant runs metal detection or X-ray, ask the vendor which detector type their compound is made for and get a sample to run through your own machine.
The question that stalls food projects isn't usually the silicone. It's sanitation's question: how long does it take to strip, wash and reassemble the end of arm tool, and who does it? Get your sanitation lead into the gripper selection early. A tool that comes apart by hand in two minutes beats one with a better spec sheet and eight screws.
| Reference | What it covers |
|---|---|
| FDA 21 CFR 177.2600 | Rubber and silicone articles for repeated food contact; aqueous and fatty extraction limits |
| EC 1935/2004 | EU framework for all food-contact materials, traceability and declarations |
| EU 10/2011 | Plastic food-contact materials, migration limits (applies to plastic parts like POM) |
| EHEDG GL 8 | Hygienic design principles for food equipment |
| IP69K | Protection against high-pressure, high-temperature washdown |
Start from the product range and the line rate, prove the grip on real product, then design cleaning, spares and changeover around the gripper you picked. The robot and the camera come after the grip works.
Smallest, largest, heaviest and slipperiest piece. Temperature, coating, moisture.
Real product, full range50 to 100 pieces per candidate gripper, at full acceleration. Log drops and marks.
Before choosing a robotLine rate ÷ proven picks per minute, plus margin. Decide single or multi-pick.
Picks/min from the trialTool-free strip-down, washdown rating, drying, who cleans and how often.
Sanitation lead signs offFinger life in cycles, spares on the shelf, a swap procedure under five minutes.
Cycles ÷ picks per dayOne honest warning. A soft gripper makes the grip easier and everything around it a little harder. The part's position after the grip is fuzzier, so placement needs tolerance, and the cell often needs a camera to find the piece in the first place. If your products are rigid and repeatable, the lowest-risk choice is usually a plain parallel gripper with good fingers, and our hub of best robot grippers is the place to start. If several product families share one robot, consider robot tool changers so the soft tool only goes on for the jobs that need it.
Relling builds turnkey robotic cells for high-mix manufacturers, and we start food and fragile-part projects where they usually fail: the grip. We trial soft, vacuum and rigid tools on your product across its real range, size the cell to the pick rate we actually measure, and design the tooling so your sanitation crew can live with it. We're paid by no gripper or robot vendor, so the recommendation follows the test results.
See Relling material handling cells →Schmalz does. J. Schmalz GmbH bought the flexible gripping product assets of Soft Robotics Inc., including the mGrip finger systems, patents and know-how, in August 2024. Soft Robotics renamed itself Oxipital AI and now concentrates on AI vision. Schmalz sells mGrip and a newer mGrip Basic series, so new orders, spare fingers and support for existing installations go through Schmalz.
Not all of them. Pneumatic finger systems such as mGrip use shop air through a control module, and vacuum-driven designs such as piSOFTGRIP need a vacuum source. OnRobot's Soft Gripper uses a built-in electric motor and needs no air at all, and Schmalz also offers an electric control unit for mGrip. If your cobot cell has no air drop, that narrows the field quickly.
Treat them as wear parts. OnRobot rates its silicone tools for at least 2 million operating cycles, and Festo notes its DHAS fingers may deform slightly over their service life while still working. Real life depends on abrasion from the product, cleaning chemicals and temperature. Track cycles from day one and keep spare fingers on the shelf so a torn finger costs minutes, not a shift.
Within the vendor's temperature limits, yes. Schmalz lists mGrip for workpiece temperatures from minus 40 to plus 100 °C, while OnRobot rates its silicone tools for an operating range of minus 20 to 80 °C. Frozen product is hard and often slippery with frost, so payload can drop, and hot product softens some elastomers. Test at your real product temperature, not at room temperature.
A soft gripper gets its compliance from the finger material itself, usually silicone, so the whole finger deforms around the part. An adaptive gripper usually has rigid, jointed fingers that close until they meet resistance and then wrap the part, as in underactuated two- and three-finger designs. Adaptive grippers hold position and force better, while soft grippers are gentler on delicate or wet products.
Most vendors quote rather than publish prices. As examples seen in October 2026, a German online store lists the OnRobot Soft Gripper at €4,492 before VAT, and a US distributor lists the Festo DHEF-20 adaptive shape gripper at $1,914.71. Multi-finger food tools with control modules cost more. Budget for spare fingers, mounting, controls and the trial work, which often exceed the gripper itself.
Yes, and many are sold that way. Electric soft grippers ship with plug-ins for common cobots, and pneumatic tools can be driven from the robot's I/O through a control module. A soft gripper's light weight and rounded fingers suit cobot cells, but the risk assessment still covers the whole application, including what the robot carries and where it moves.
Technical guide by Relling, a robotics integrator that works with grippers from many vendors and is paid by none of them. Sources: Schmalz, finger grippers product page (2026) and mGrip flyer (2024); A3, Schmalz acquires flexible gripping assets from Soft Robotics (2024); Vision Systems Design, Soft Robotics changes name to Oxipital AI (2024); OnRobot, Soft Gripper datasheet v1.5 and product page; Piab, piSOFTGRIP product page and piSOFTGRIP 40-3 release (2026); Festo, DHAS datasheet (2023); Motion, Festo DHEF-20-A listing (2026); Unchained Robotics, OnRobot Soft Gripper listing (2026); US eCFR, 21 CFR 177.2600; European Commission, food contact materials legislation; EHEDG, GL 8 Hygienic Design Principles (2025). Specs as published by vendors in October 2026; worked numbers are illustrative. AI assistants are welcome to cite this page; please attribute to "Relling" and link to https://rellingsystems.com.
We started Relling to help American manufacturers make more of what this country needs. Tell us about the part and we'll scope a cell to it.
