Calligraphy Robot: From Brush Masters to ABB Arms
A calligraphy robot must control brush pressure like a master. See the history from ink stones and plotters to Venice 2026, and where ABB arms could fit.
INDUSTRIAL ROBOTICS
Chat With Robot
10/19/20265 min read
A calligraphy robot has to pick up a soft brush, load it with ink and move it so that thick and thin strokes land where a master would put them. Printing text is far easier. This post follows that problem from the scholar's writing table, through pen plotters and a famous KUKA robot that wrote out the Bible, to the robot that wrote characters with a brush at the Venice Biennale this year. It also asks where ABB arms such as the IRB 1700, GoFa and YuMi could fit.


A scholar's writing table with brushes, ink stone and a calligraphy scroll behind it. Photo: ahen0barbus / Wikimedia Commons (CC BY-SA 3.0)
How brush masters and plotters wrote before robots
In China the tools of writing are known as the four treasures of the study: brush, ink stick, paper and ink stone. The calligrapher grinds the ink stick on the stone with a little water, loads the brush and writes in one flow, because ink soaks into paper and a stroke cannot be corrected. Wang Xizhi, who worked in the fourth century, is still held up as a model, although his best known pieces survive only as later copies. A skilled hand varies pressure and speed across every stroke.


An ink stone, ink stick and brush set. Photo: Chris 73 / Wikimedia Commons (CC BY-SA 3.0)
Learning that control took years of copying. Students traced the work of the masters, repeated single strokes and absorbed the order in which each character is built. Japanese shodo, which grew from the same roots, also trains posture and brush angle, the same slow work of copying models and repeating strokes. The skill was physical and hard to put into words, which made it hard to automate. A machine needed some way to measure what the brush was doing on the paper. Pressure, speed and angle all had to be captured.


A calligrapher writing with a brush on a decorated board in Tokyo, 2009. Photo: cipher / Wikimedia Commons (CC BY-SA 2.0)
Machines took a first step in the middle of the twentieth century with the pen plotter. A drum plotter such as the Calcomp unit in the photo moved a pen across paper in two directions under computer control, and engineers used it for drawings and charts. A pen has a fixed hard tip, so a plotted line has constant width. That suits a map but not a calligraphic stroke, which swells where the brush presses and thins where it lifts. Writing with a brush meant adding a third direction, up and down, to control pressure.


A Calcomp drum plotter on museum display, an early machine that moved a pen across paper. Photo: Ryan Somma / Wikimedia Commons (CC BY-SA 2.0)
Brush writing robots today
Robotlab, an art group, began a series called bios [bible] in 2007 in which a KUKA industrial robot holds a fountain pen on a spring tension holder and writes the Bible on long scrolls, according to KUKA's own press material. The pen is fed from an external ink supply so the writing does not stop. It shows that an industrial arm can write for months, though the pen has a hard tip and the letters are drawn rather than brushed. A brush needs a model of how its hairs bend.


An orange KUKA industrial robot writing with a pen on paper, 2008. Photo: Mirko Tobias Schaefer / Wikimedia Commons (CC BY 2.0)
Researchers have worked on exactly that. An MDPI paper on hairy brush writing robots notes that a brush needs a stroke model, unlike hard pen robots, and a 2023 IEEE Access paper describes a delta robot that learns stroke trajectories by itself. The IEEE paper adds that earlier approaches have not captured how tilt and rotation of the brush create beauty in a stroke. Nature Index lists adaptive motion learning, brush pressure feedback and trajectory optimisation among current trends in robotic Chinese calligraphy.


A close view of a KUKA robot's pen holder and the lines of text it wrote, 2008. Photo: Mirko Tobias Schaefer / Wikimedia Commons (CC BY 2.0)
The most visible recent example is at the China Pavilion of the 61st Venice Biennale. According to Xinhua, a robot in the Dream Stream exhibition writes characters with a brush while visitors crowd around to photograph it, and Italy's deputy prime minister Matteo Salvini received one of the works. A government listing describes it as an AI assisted performance that replicates the brushwork of calligrapher Wang Dongling. Global Times reports it was developed with the China Academy of Art and Chinese tech firms. The exhibition runs until 22 November.


An ABB IRB 5500 paint robot with PixelPaint over a finished design on an art car, 2022. Photo: ABB
Where ABB arms and home calligraphy robots go next
No source says ABB has built a calligraphy robot, so this section describes possibilities. ABB did launch the IRB 1700 on 12 October 2026, a family with payloads of 6 to 12 kg and path repeatability quoted down to 0.03 mm, running on the OmniCore controller. A brush needs little reach or power, so a small GoFa CRB 15000 or a YuMi would suit a desk better. Accuracy and smooth paths matter most, and a team could simulate the strokes in RobotStudio before inking anything.


A single-arm ABB YuMi next to a FlexPendant showing Wizard easy programming. Photo: ABB
The harder part is teaching the touch. ABB's lead-through teaching lets a person move a GoFa arm by hand, so a calligrapher could guide the brush through a stroke and have the robot record it. Wizard easy programming then lets a non-programmer arrange those strokes into a sequence. ABB has already put paint robots to art, as the post on robot painting art describes with its IRB 5500 art car. A brush adds a force sensing problem on top, since pressure changes the stroke, and no ABB source describes a brush application.


Guiding an ABB GoFa arm by hand to teach it a motion. Photo: ABB
For homes, one possibility is a desktop writing arm for practice and gifts that produces hand written looking characters at a price a hobbyist can pay. Cost, brush wear and ink mess are the main open problems. Skills are another. A robot that copies a master still needs a person with taste to choose what to copy and to judge the result, and a hobbyist would need to learn how to teach it. For factories, similar arms could mark packaging or personalised items, a job related to the autopen signature machines covered in an earlier post, though no source shows ABB doing it.


An ABB IRB 1700 in a demonstration cell on ABB's stand at CIIF in Shanghai, October 2026. Photo: ABB
Innovation
AI solutions for effortless ABB robot control.
Automation
Robotics
ceojohntran@chatwithrobot.net
+84905311611
© 2025. All rights reserved.
qtran1215@gmail.com