TIAGo (Take It And Go) is a modular mobile manipulator service robot developed by PAL Robotics, a Spanish robotics company headquartered in Barcelona. The platform was first introduced in February 2015 and is designed as a customisable research and service robot, available in single-arm, dual-arm, and base-only configurations, with options for a lifting torso, a head with display, and a wide range of end-effectors. TIAGo runs on the Robot Operating System, ships with a Gazebo simulation model, and has become one of the most widely used mobile manipulation platforms in European academic robotics research and in commercial deployments across healthcare, retail, hospitality, and assistive applications.
Unlike many of the humanoid platforms in PAL Robotics' lineup, such as REEM-C, TALOS, and Kangaroo, TIAGo is wheeled rather than bipedal. This makes it cheaper to operate, safer in human environments, and more practical for everyday service tasks like fetching objects, carrying trays, or scanning shelves. Over more than a decade, the family has expanded into a full product line that includes the standard TIAGo, the dual-arm TIAGo++, the omnidirectional TIAGo OMNI Base, the inventory-focused StockBot, and the more recent TIAGo Pro Edition with series elastic actuator (SEA) arms.
TIAGo is a wheeled mobile manipulator: a robot that combines a mobile base for indoor navigation with a multi-degree-of-freedom arm for object manipulation. The standard configuration stands roughly 1.10 to 1.45 metres tall thanks to a telescoping torso, weighs around 70 kilograms, and uses a differential-drive base with an integrated laser scanner for navigation. A 7 degree-of-freedom arm carries up to 3 kilograms of payload, and a 2 DoF pan-tilt head holds an RGB-D camera for perception. The robot can be ordered with no arm, one arm, or two arms, with several gripper options, and with computing hardware ranging from an Intel i5 to an Intel i7 with optional NVIDIA Jetson GPU integration.
The name TIAGo, an acronym for Take It And Go, reflects the robot's intended role: pick things up, move them around, and hand them to people. It belongs to the same generation of service-oriented mobile manipulators as the Toyota Human Support Robot, the now-discontinued Fetch, and the smaller Hello Robot Stretch. Among these, TIAGo is the most modular: customers choose almost every major subsystem, including the drive type, the laser model, the arm count, the actuation technology, the gripper, and the optional accessories such as force/torque sensors, touchscreens, additional batteries, and Velodyne LiDAR mounts.
PAL Robotics was founded in Barcelona in 2004 to develop service and research robots for European customers. Francesco Ferro, who serves as CEO, is one of the company's co-founders, and longtime CTO Luca Marchionni has led much of its technical strategy. The company today operates from Barcelona with additional teams in Toulouse and Italy, and ships robots to customers in roughly fifteen countries. Its first product, the REEM-A humanoid, grew out of an early project to build a robotic arm that could play chess.
TIAGo arrived in 2015 as PAL Robotics' first dedicated mobile manipulator. By that point the company had already shipped three generations of REEM humanoids and was about to launch the REEM-C research humanoid. TIAGo was designed to bring the company's expertise in humanoid arms and perception to a wheeled, more practical form factor that universities and integrators could afford. It was first announced on the PAL Robotics blog and covered by IEEE Spectrum in February 2015 with three trim levels nicknamed "iron," "steel," and "titanium," priced from roughly 30,000 to 60,000 euros.
| Year | Milestone |
|---|---|
| 2004 | PAL Robotics founded in Barcelona, Spain. Francesco Ferro is co-founder and CEO. |
| 2006 | REEM-A humanoid demonstrated as a chess-playing robotic arm project that grew into a full humanoid. |
| 2010 | REEM (third-generation REEM, sometimes called REEM-H) introduced for hospitality and public spaces. |
| 2013 | REEM-C full-size biped humanoid announced as a research platform with 44 degrees of freedom. |
| 2015 | TIAGo launched in February with three trim levels (iron, steel, titanium). |
| 2017 | TALOS humanoid introduced jointly with LAAS-CNRS in Toulouse, with torque-controlled joints. |
| 2017 | Co4Robots Horizon 2020 project starts; PAL Robotics joins as a TIAGo provider. |
| 2017 | StockBot deployed in retail (Decathlon Singapore was an early customer in 2019). |
| 2019 | TIAGo++ dual-arm version released for bimanual research. |
| 2020 | TIAGo OMNI Base released with a mecanum-wheel omnidirectional drive. |
| 2022 | ARI social robot launched after pilots in residential care. |
| 2022 | Kangaroo bipedal humanoid presented at ICRA 2022 in Philadelphia. |
| 2023 | TIAGo Pro Edition unveiled at ICRA 2023 in London (29 May to 2 June 2023). |
| 2024 | TIAGo SEA arm announced as a customisation option, developed in the EU CANOPIES project. |
| 2024 | TIAGo OMNI++ released, combining the OMNI Base with two SEA arms. |
| 2025 onward | Continued integration with the LeRobot ecosystem and other foundation-model toolchains. |
TIAGo is sold as a family of robots that share a common software stack, a common look, and many shared components, but differ in arm count, base type, and target market. The variants below cover the main commercial offerings as of 2024 and 2025.
| Variant | Arms | Base | Target use | Notes |
|---|---|---|---|---|
| TIAGo Base | 0 | Differential or omni | Logistics, navigation research | The mobile base on its own; can be upgraded later. |
| TIAGo (standard) | 1 (7-DoF) | Differential | Research, service tasks | The original 2015 configuration. |
| TIAGo Steel | 1 (7-DoF) | Differential | Research with light manipulation | Mid-tier 2015 trim, standard parallel gripper. |
| TIAGo Titanium | 1 (7-DoF) | Differential | Advanced research | High-end 2015 trim with five-fingered hand and force/torque sensor. |
| TIAGo Iron | 0 | Differential | Mobile-only research | Base 2015 trim with 3 m laser, no arm. |
| TIAGo++ | 2 (7-DoF each) | Differential | Bimanual research | Each arm reaches 92 cm; both reach the floor and up to 1.75 m. |
| TIAGo OMNI Base | 0 | Omnidirectional (mecanum) | Intralogistics, retail | 360 degree FOV from two LiDARs, web fleet UI. |
| TIAGo OMNI++ | 2 (SEA, 7-DoF each) | Omnidirectional | Industry, advanced research | Combines OMNI Base with two SEA arms. |
| TIAGo Pro Edition | 1 or 2 (SEA, 7-DoF) | Omnidirectional (mecanum) | Industry and human-robot collaboration | Unveiled at ICRA 2023; new arm placement, torque control, ROS 2. |
| TIAGo Lift | 1 | Differential | Pick from low and high shelves | Telescoping torso (1.10 to 1.45 m). |
| TIAGo Stockbot (StockBot) | 0 | Differential | Retail inventory | RFID antennas, deployed by Decathlon worldwide. |
| TIAGo Junior | 1 | Differential | Education, workshops | Smaller, simplified academic configuration. |
End-effectors available across the line include a parallel gripper, the Robotiq 2F-85 and 2F-140 grippers, the Robotiq Hand-E, an ePick vacuum gripper, PAL Robotics' own PAL gripper, and a five-fingered Schunk-style hand on the older Titanium trim. Most grippers can be ordered with an integrated 6-axis force/torque sensor at the wrist.
TIAGo's hardware reflects its dual identity as both a research platform and a service robot. The standard TIAGo, sold continuously from 2015, has the following core specifications. Numbers are drawn from the public PAL Robotics datasheets and the IEEE Spectrum launch coverage.
| Component | Specification |
|---|---|
| Height | 1.10 to 1.45 m (telescoping torso) |
| Footprint | Approximately 54 cm diameter |
| Weight | About 70 kg (standard); 96 kg (TIAGo Pro) |
| Maximum speed | 1.0 to 1.5 m/s |
| Mobile base (standard) | Differential drive with two driven wheels and casters |
| Mobile base (OMNI) | Four mecanum wheels, omnidirectional |
| Navigation laser | Hokuyo URG-04LX (5.6 m) or SICK TiM561 (10 m or 25 m) |
| Arm | 7 DoF, optional Series Elastic Actuator (SEA) variant |
| Arm payload | 3 kg (without end-effector) |
| Arm reach | Up to 96 cm with the SEA arm; about 87 cm with standard arm |
| Head | 2 DoF pan-tilt with RGB-D camera (ASUS Xtion or Intel RealSense) |
| Wrist sensor | Optional 6-axis force/torque sensor |
| Speakers and microphones | Stereo speakers, dual microphones |
| Battery | 36 V, 20 Ah; 4 to 5 hours single battery, 8 to 10 hours dual |
| Onboard computer | Intel i5/i7, 8 to 16 GB RAM, 250 to 500 GB SSD |
| GPU option | NVIDIA Jetson TX2 or newer Jetson modules |
| Operating system | Ubuntu LTS with a real-time kernel |
| Middleware | ROS 1 (Melodic, Noetic) and ROS 2 |
The Pro Edition uses different numbers in several places. Its omnidirectional base is faster (around 1.5 m/s), it weighs about 96 kg, and its torque-controlled SEA arms each carry 3 kg with a 96 cm reach. The Pro Edition also adds two long-range LiDAR sensors for full 360 degree coverage, a face screen with LEDs for expressive gaze, and offline wake-word voice processing.
TIAGo runs on the Robot Operating System out of the box, and it is one of the few commercial mobile manipulators with first-party support in both ROS 1 and ROS 2. PAL Robotics maintains official ROS packages on GitHub at pal-robotics/tiago_robot, pal-robotics/tiago_dual_robot, and pal-robotics/tiago_simulation, and the platform has its own page on robots.ros.org. The simulation packages let researchers run a complete TIAGo model in Gazebo before touching the real hardware, which is useful for rapid prototyping, education, and reproducible benchmarks.
On top of vanilla ROS, PAL Robotics ships its own software stack called PAL OS. Premium add-on packages include Whole Body Control for joint-space and Cartesian motion, Advanced Navigation for mapping and obstacle avoidance, Advanced Grasping based on behaviour trees with computer-vision integration, an Animation Interface for synchronised motion, speech, and LED effects, and a Visual Programming environment that lets non-programmers build task workflows. Web-based teleoperation lets operators drive the robot, look through its head camera, and trigger pre-programmed routines from a browser. Standardised APIs make higher-level integration straightforward, and PAL provides reference applications for hand-over, hand-shake, navigation, and pick-and-place tasks.
In 2025, the LeRobot project from Hugging Face added support for TIAGo, opening the platform to imitation-learning and foundation-model workflows used widely across the open robotics community. Combined with the Pro Edition's torque-controlled arms, this makes TIAGo a credible host for learning-from-demonstration research.
TIAGo has been deployed in a long list of application domains. The list below is grouped by sector and cites the specific PAL Robotics deployments or EU project use cases that drive the application.
PAL Robotics has been a long-standing partner in European Union framework projects, and TIAGo is the most common platform it brings to those collaborations. The list below covers some of the well-documented programmes.
| Project | Period | Focus |
|---|---|---|
| Co4Robots | 2017 to 2020 | Decentralised control of multi-robot teams; PAL contributed TIAGo and TIAGo Base. |
| MoveCare | 2017 to 2020 | Assistive multi-user platform for elderly people living alone, integrating a Social Assistive Robot. |
| ENRICHME | 2015 to 2018 | Long-term human-robot interaction for older adults at home; uses TIAGo. |
| GrowMeUp | 2015 to 2018 | Companion robot for active ageing in Europe, using TIAGo as the platform. |
| SHAPES | 2019 to 2023 | Smart and healthy ageing through people engaging in supportive systems; ARI and TIAGo pilots in residential care. |
| CANOPIES | 2021 to 2025 | Robots collaborating with humans in vineyards; used TIAGo with the new SEA arm. |
| AGIMUS | 2022 to 2026 | Agile manufacturing using collaborative robots, including TIAGo Pro. |
| PILLAR-Robots | 2022 to 2026 | Purposeful intrinsically-motivated learning architecture for robots; uses TIAGo for use-case demonstration. |
| euROBIN | 2022 to 2026 | European network on AI-powered robotics; multiple TIAGo and TIAGo++ entries in the service-robot coopetition. |
The MURAB project, often listed in mobile-manipulation reviews, was led by the University of Twente with KUKA Robotics and Siemens; it used a different robotic arm, not a TIAGo, although it sits in the same European medical-robotics ecosystem and is sometimes grouped with PAL's healthcare work.
TIAGo competes with, and is regularly compared against, a small number of mobile-manipulation platforms aimed at research labs and service applications. The table below sketches the main alternatives. "ROS support" indicates first-party rather than community packages.
| Platform | Maker | Year | Form factor | Arms | Height | Payload | ROS support | Focus |
|---|---|---|---|---|---|---|---|---|
| TIAGo | PAL Robotics | 2015 | Wheeled mobile manipulator | 0, 1, or 2 (7-DoF) | 1.10 to 1.45 m | 3 kg per arm | ROS 1 and ROS 2 | Modular research and service |
| Toyota HSR | Toyota | 2015 | Wheeled, compact | 1 (5-DoF folding) | 1.00 to 1.35 m | 1.2 kg | ROS 1 | Compact home assistance, RoboCup@Home |
| Fetch | Fetch Robotics | 2015 | Wheeled mobile manipulator | 1 (7-DoF) | About 1.50 m | 6 kg | ROS 1 | Research and warehousing; discontinued circa 2020 after Zebra acquisition |
| Stretch (RE1, RE2, RE3) | Hello Robot | 2020 | Wheeled, lightweight | 1 (telescoping Cartesian) | About 1.40 m | About 1.5 kg | ROS 1 and ROS 2 | Compact, low-cost (~25,000 USD) |
| Robotnik RB-1 / RB-2 | Robotis (separate vendor: Robotnik) | 2014 onward | Modular base + arm | 0 to 2 | Configurable | Configurable | ROS 1 and ROS 2 | Modular, often paired with UR arms |
| Kinova Movo | Kinova | 2017 | Wheeled mobile manipulator | 1 or 2 (Kinova arms) | About 1.40 m | 2 kg per arm | ROS 1 | Research; discontinued |
| Cobalt | Cobalt Robotics | 2017 | Wheeled, no arm | 0 | About 1.50 m | n/a | Proprietary | Indoor security patrol |
| Universal Robots e-Series | Universal Robots | 2018 | Industrial cobot arm | 1 | n/a | 3 to 16 kg | ROS 1 and ROS 2 | Cobot arm only, not mobile |
| Spot + arm | Boston Dynamics | 2020 | Quadruped + 6-DoF arm | 1 | About 0.80 m | 11 kg | ROS 1 (community) | Inspection, manipulation in unstructured terrain |
| Apptronik Apollo | Apptronik | 2023 | Bipedal humanoid | 2 | About 1.73 m | 25 kg per arm | Proprietary | General-purpose humanoid |
| 1X NEO | 1X Technologies | 2024 | Bipedal humanoid | 2 | 1.65 m | About 20 kg total | Proprietary | Home assistant humanoid |
Compared with the Toyota HSR, TIAGo is bigger, has a longer arm reach, and is more open to customisation. HSR is more compact and more standardised, which is why it remains the default platform for the RoboCup@Home benchmark league. Compared with Hello Robot Stretch, TIAGo is significantly more expensive but offers higher payload, multi-arm configurations, and an industrial-grade build. Compared with the discontinued Fetch, TIAGo benefits from PAL Robotics' continued support and new variants, including the Pro Edition.
TIAGo is one of the most-used mobile-manipulation platforms in European robotics research. It often appears alongside the Toyota Human Support Robot in academic comparisons of service-robot platforms, and PAL Robotics reports more than 40 collaborative projects, deployments in 14 EU countries, and over 120 academic citations on the platform's product page. The robot's modularity is its main selling point: a customer can buy a base today, add an arm later, swap a gripper, or upgrade a laser without replacing the whole machine.
For PAL Robotics, TIAGo is a flagship product that connects the company's humanoid heritage (REEM, REEM-C, TALOS, Kangaroo) with its newer push into commercial service robotics (StockBot, ARI). The company is one of the few European robotics manufacturers with a full product line spanning humanoids, mobile manipulators, and social robots. CEO Francesco Ferro has been outspoken about Europe's bureaucratic and funding challenges in robotics relative to North America and Asia, and PAL's continued shipment of TIAGo to research and industrial customers is one of the strongest pieces of evidence that European service-robotics manufacturers can compete at the high end.
The April 2025 acquisition of Pollen Robotics (the maker of the Reachy 2 dual-arm robot) by Hugging Face brought renewed attention to the broader European service-robot ecosystem, of which TIAGo is the most established member. The integration of TIAGo with the open-source LeRobot toolchain makes it a natural target for foundation-model research and learning-from-demonstration experiments in 2025 and beyond.
For a wiki summary, it helps to be honest about what TIAGo does well and where it falls short. The list below pulls together points from PAL's own marketing, third-party reviews on Robots Guide and IEEE Spectrum, and the academic literature.
Strengths
Limitations
Since 2023, PAL Robotics has spent most of its TIAGo road map on three threads: rolling out the Pro Edition with SEA arms, broadening the OMNI base into a full mobile-manipulation product (TIAGo OMNI++), and tying the platform into the broader open foundation-model ecosystem.
The SEA arm, first announced as part of the EU CANOPIES vineyard project, became a customisation option in 2024. It offers torque-controlled compliance, a 96 cm reach, and the dexterity needed for tasks like grape pruning that require gentle, force-aware contact. The TIAGo OMNI++, released in 2024, combines the omnidirectional mecanum-wheel base with two SEA arms, a face screen, and ROS 2 support. PAL Robotics has indicated continued investment in this configuration through ICRA 2026 and beyond.
On the software side, the addition of TIAGo to the LeRobot project supports imitation learning and foundation-model workflows. PAL Robotics also continues to take part in EU-funded foundation-model and AI-robotics initiatives such as PILLAR-Robots and euROBIN, both of which use TIAGo and TIAGo++ as host platforms. Healthcare deployments continue to grow: the SHAPES project's residential-care pilots expanded after 2023, and the ARI social robot, which shares mobile-base technology with TIAGo, complements the line-up where manipulation is not required.