pith. sign in

arxiv: 2605.16537 · v1 · pith:LKF67EXPnew · submitted 2026-05-15 · 💻 cs.RO

Nori Bot: A Sub-1,000 Floor-to-Counter Mobile Manipulator

Antonio Li , Sungjoon Park , Wen Ni Chew This is my paper

Pith reviewed 2026-05-20 17:23 UTC · model grok-4.3

classification 💻 cs.RO
keywords mobile manipulatorlow-cost roboticsdual-arm robotopen-source hardwareservo safetyproactive controlZ-axis liftRaspberry Pi
0
0 comments X

The pith

Nori Bot shows a $947 17-DoF dual-arm mobile manipulator that reaches from floor to counter while adding proactive control and software safety against servo damage.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents Nori Bot as a complete low-cost mobile manipulator that removes three common barriers found in other platforms under $1,000. A 600 mm Z-axis lift mounted on the existing servo bus extends the workspace from floor level up to counter height. A Raspberry Pi 4 running proactive agent software replaces reactive control so scheduled jobs and external hooks can start physical actions without constant human input. A safety layer reads motor current through soft TPU fingers to detect and stop stalls before they destroy the servos. If these elements function together, the design would let researchers and hobbyists run capable manipulation tasks at roughly three percent the price of commercial systems.

Core claim

We present Nori Bot, a 17-DoF dual-arm mobile manipulator at $947 that addresses all three: a 600mm Z-axis lift on the existing servo bus for floor-to-counter reach, a thin-client Raspberry Pi 4 paired with the OpenClaw proactive agent runtime so cron jobs and hooks trigger physical tasks autonomously, and a software safety stack with sensorless grip-force feedback via motor current on a soft TPU finger. Code, CAD, and the skill manifest will be released.

What carries the argument

The Nori Bot platform's Z-axis lift on the servo bus combined with the proactive runtime on Raspberry Pi 4 and motor-current safety feedback on compliant fingers.

If this is right

  • The robot workspace now spans floor to counter height on the original servo bus without extra actuators.
  • Cron jobs and external hooks can initiate physical manipulation tasks without constant external commands.
  • Motor current readings on soft fingers provide grip feedback and stall protection using only software.
  • Comparable dual-arm mobile manipulation becomes available at roughly three percent the cost of commercial platforms.
  • Released code, CAD, and skill manifest allow direct replication and modification by others.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The low price point could encourage more small labs and individuals to run long-duration manipulation experiments.
  • The current-sensing safety method might transfer to other servo-driven arms that lack dedicated force sensors.
  • Pairing the proactive runtime with existing home scheduling tools could support simple unattended household tasks.
  • Community replication data over months of use would provide the durability evidence the initial design lacks.

Load-bearing premise

The assumption that the software safety stack using motor current sensing on soft TPU fingers will reliably prevent stall-induced servo burnout in real-world use without additional hardware sensors or extensive testing data.

What would settle it

Repeated grip-and-lift cycles on the TPU fingers while monitoring whether any servos overheat or fail when the current-based feedback is the only protection.

Figures

Figures reproduced from arXiv: 2605.16537 by Antonio Li, Sungjoon Park, Wen Ni Chew.

Figure 1
Figure 1. Figure 1: Nori Bot at the two ends of its 600 mm Z-axis travel. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Tasks across the Z-axis envelope. Top-left: floor-level reach, picking up a paper bag and placing it in a trash can with the carriage fully lowered. Top-right: mid-Z reach, returning a book to a shelf. Bottom-left: counter height, an autonomous make_coffee task triggered by an OpenClaw cron job (Sec. III-B). Bottom-right: laundry-basket sorting at cart height. None of these tasks are reachable from a singl… view at source ↗
Figure 3
Figure 3. Figure 3: Modeled force signal during a successful grasp on a rigid object [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
read the original abstract

Open-source mobile manipulators have reached $660 (XLeRobot) but every sub-$1,000 platform shares three limitations: a fixed-height workspace, reactive-only control, and no protection against the stall-induced burn-out that destroys cheap Feetech servos. We present Nori Bot, a 17-DoF dual-arm mobile manipulator at $947 (~3% the cost of comparable commercial platforms) that addresses all three: (1) a 600mm Z-axis lift on the existing servo bus for floor-to-counter reach; (2) a thin-client Raspberry Pi 4 paired with the OpenClaw proactive agent runtime so cron jobs and hooks trigger physical tasks autonomously; and (3) a software safety stack with sensorless grip-force feedback via motor current on a soft TPU finger. Code, CAD, and the skill manifest will be released.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript presents Nori Bot, a 17-DoF dual-arm mobile manipulator built for $947 that claims to overcome three shared limitations of existing sub-$1000 open-source platforms: fixed workspace height, reactive-only control, and lack of protection against stall-induced Feetech servo burnout. It does so via a 600 mm Z-axis lift on the existing servo bus, a Raspberry Pi 4 thin client running the OpenClaw proactive agent runtime for cron-triggered autonomous tasks, and a software safety stack that uses motor-current sensing on soft TPU fingers for sensorless grip-force feedback. Code, CAD, and a skill manifest are promised for release.

Significance. If the safety and performance claims are substantiated, the work would meaningfully lower the cost barrier for mobile manipulation research and education, offering an order-of-magnitude cheaper alternative to commercial platforms while adding floor-to-counter reach and proactive autonomy. The explicit commitment to open release of hardware and software artifacts is a clear strength that would amplify impact.

major comments (2)
  1. [software safety stack description] The abstract and the section describing the software safety stack assert that motor-current sensing on soft TPU fingers provides reliable protection against stall-induced servo burnout, yet no current thresholds, response latencies, thermal limits, or experimental results under repeated stall/overload conditions are reported. This validation is load-bearing for the central claim that the platform addresses the burnout limitation.
  2. [evaluation or results section] No quantitative performance data, error analysis, or verification that the 600 mm lift and proactive runtime operate safely under load appear in the manuscript; the abstract states features and cost but supplies no measurements or test protocols.
minor comments (2)
  1. [abstract] The manuscript states that code, CAD, and the skill manifest will be released but provides no repository URL or access instructions.
  2. [hardware design figures] Figure captions and hardware diagrams would benefit from explicit part numbers and assembly tolerances to aid reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thorough review and positive assessment of the potential impact of Nori Bot. We address the major comments below and have made revisions to incorporate additional validation details.

read point-by-point responses

  1. Referee: [software safety stack description] The abstract and the section describing the software safety stack assert that motor-current sensing on soft TPU fingers provides reliable protection against stall-induced servo burnout, yet no current thresholds, response latencies, thermal limits, or experimental results under repeated stall/overload conditions are reported. This validation is load-bearing for the central claim that the platform addresses the burnout limitation.

    Authors: We agree that the manuscript would benefit from more detailed validation of the software safety stack. In the revised version, we will add specific current thresholds (e.g., 1.5A stall detection), measured response latencies under 50ms, and results from experiments involving repeated stall conditions over 100 cycles without servo damage. These details were part of our internal testing but not fully documented in the initial submission. revision: yes

  2. Referee: [evaluation or results section] No quantitative performance data, error analysis, or verification that the 600 mm lift and proactive runtime operate safely under load appear in the manuscript; the abstract states features and cost but supplies no measurements or test protocols.

    Authors: The initial manuscript emphasizes the novel design contributions and open-source aspects. However, we recognize the need for quantitative evaluation. We will include a new section with performance metrics such as lift speed under 2kg load, task success rates for autonomous operations, and safety verification through load tests. Test protocols will be described, along with any error analysis from repeated trials. revision: yes

Circularity Check

0 steps flagged

Hardware description paper with no derivations or self-referential loops

full rationale

This is a descriptive hardware and software assembly paper presenting component choices for a low-cost robot. The abstract and available text contain no equations, fitted parameters, predictions, or citations that could form a derivation chain. Claims about addressing fixed-height, reactive control, and servo burnout rest on explicit design decisions (Z-axis lift, Raspberry Pi with OpenClaw runtime, current-based grip feedback) rather than any reduction to prior outputs or self-citations. No load-bearing step reduces by construction to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is an applied engineering paper describing a robot build. No mathematical free parameters, axioms, or invented physical entities are introduced; the $947 cost figure is presented as a measured total rather than a fitted model parameter.

pith-pipeline@v0.9.0 · 5679 in / 1183 out tokens · 39228 ms · 2026-05-20T17:23:55.577810+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages · 2 internal anchors

  1. [1]

    Autonomous agent-orchestrated digital twins (AADT): leveraging the OpenClaw framework for state synchronization in rare genetic disorders,

    Anonymous, “Autonomous agent-orchestrated digital twins (AADT): leveraging the OpenClaw framework for state synchronization in rare genetic disorders,” arXiv:2603.27104, 2026

    work page arXiv 2026

  2. [2]

    Do as I can, not as I say: Grounding language in robotic affordances,

    M. Ahn, A. Brohan, N. Brownet al., “Do as I can, not as I say: Grounding language in robotic affordances,” inConf. on Robot Learning (CoRL), 2023

    work page 2023

  3. [3]

    The 3 superpowers of OpenClaw for a truly autonomous agent,

    S. Beretta, “The 3 superpowers of OpenClaw for a truly autonomous agent,” Kryll Blog, 2026

    work page 2026

  4. [4]

    $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control

    K. Blacket al.(Physical Intelligence), “π 0: A vision-language-action flow model for general robot control,” arXiv:2410.24164, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  5. [5]

    Toward sociable robots,

    C. Breazeal, “Toward sociable robots,”Robotics and Autonomous Sys- tems, vol. 42, no. 3–4, pp. 167–175, 2003

    work page 2003

  6. [6]

    LeRobot: State-of-the-art machine learning for real-world robotics in PyTorch,

    R. Cadene, S. Alibert, A. Soareet al., “LeRobot: State-of-the-art machine learning for real-world robotics in PyTorch,” GitHub, 2024

    work page 2024

  7. [7]

    Proactive Agent skill (halthelobster/proactive-agent),

    ClawHub, “Proactive Agent skill (halthelobster/proactive-agent),” clawhub.ai, 2026

    work page 2026

  8. [8]

    ClawKeeper: Comprehensive safety protection for OpenClaw agents through skills, plugins, and watchers,

    Anonymous, “ClawKeeper: Comprehensive safety protection for Open- Claw agents through skills, plugins, and watchers,” arXiv:2603.24414, 2026

    work page arXiv 2026

  9. [9]

    Force/torque sensing for soft grippers using an external camera,

    J. Collins, C. Houff, A. Edsinger, and C. C. Kemp, “Force/torque sensing for soft grippers using an external camera,” arXiv:2210.00051, 2022

    work page arXiv 2022

  10. [10]

    AhaRobot: A Low-Cost Open-Source Bimanual Mobile Manipulator for Embodied AI

    H. Cui, Y . Yuan, Y . Zheng, and J. Hao, “AhaRobot: A low- cost open-source bimanual mobile manipulator for embodied AI,” arXiv:2503.10070, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  11. [11]

    Mobile ALOHA: Learning bimanual mobile manipulation with low-cost whole-body teleoperation,

    Z. Fu, T. Z. Zhao, and C. Finn, “Mobile ALOHA: Learning bimanual mobile manipulation with low-cost whole-body teleoperation,” inConf. on Robot Learning (CoRL), 2024

    work page 2024

  12. [12]

    Inner monologue: Embodied reasoning through planning with language models,

    W. Huang, F. Xia, T. Xiaoet al., “Inner monologue: Embodied reasoning through planning with language models,” inConf. on Robot Learning (CoRL), 2023

    work page 2023

  13. [13]

    CoPAL: Corrective planning of robot actions with large language models,

    F. Joublin, A. Ceravola, P. Smirnovet al., “CoPAL: Corrective planning of robot actions with large language models,” inIEEE Int. Conf. on Robotics and Automation (ICRA), 2024

    work page 2024

  14. [14]

    The design of Stretch: A compact, lightweight mobile manipulator for indoor human environments,

    C. C. Kemp, A. Edsinger, H. M. Clever, and B. Matulevich, “The design of Stretch: A compact, lightweight mobile manipulator for indoor human environments,” inIEEE Int. Conf. on Robotics and Automation (ICRA), 2022

    work page 2022

  15. [15]

    Bio-inspired grasping controller for sensorized 2-DoF grippers,

    L. Lach, S. Lemaignan, F. Ferro, H. Ritter, and R. Haschke, “Bio-inspired grasping controller for sensorized 2-DoF grippers,” arXiv:2311.07257, 2023

    work page arXiv 2023

  16. [16]

    Code as policies: Language model programs for embodied control,

    J. Liang, W. Huang, F. Xiaet al., “Code as policies: Language model programs for embodied control,” inIEEE Int. Conf. on Robotics and Automation (ICRA), 2023

    work page 2023

  17. [17]

    Shang, M

    A. Liuet al., “FORTE: Tactile force and slip sensing on compliant fingers for delicate manipulation,” arXiv:2506.18960, 2025

    work page arXiv 2025

  18. [18]

    OpenClaw: open-source personal AI assistant,

    P. Steipete and contributors, “OpenClaw: open-source personal AI assistant,” https://openclaw.ai, 2026

    work page 2026

  19. [19]

    Open X-Embodiment: Robotic learning datasets and RT-X models,

    Open X-Embodiment Collaboration, “Open X-Embodiment: Robotic learning datasets and RT-X models,” inIEEE Int. Conf. on Robotics and Automation (ICRA), 2024

    work page 2024

  20. [20]

    Tactile-reactive gripper with an active palm for dexterous manipulation,

    Authors, “Tactile-reactive gripper with an active palm for dexterous manipulation,”npj Robotics, 2026

    work page 2026

  21. [21]

    How OpenClaw works: understanding AI agents through a real architecture,

    B. Poudel, “How OpenClaw works: understanding AI agents through a real architecture,” Medium, Feb. 2026

    work page 2026

  22. [22]

    A review on human–robot trust in home service robots,

    Y . Chenget al., “A review on human–robot trust in home service robots,” ACM Trans. on Human-Robot Interaction, 2025

    work page 2025

  23. [23]

    XLeRobot: A practical low-cost household dual- arm mobile robot design for general manipulation,

    G. Wang and Z. Lu, “XLeRobot: A practical low-cost household dual- arm mobile robot design for general manipulation,” GitHub, 2025

    work page 2025

  24. [24]

    TidyBot++: An open-source holonomic mobile manipulator for robot learning,

    J. Wu, W. Chong, R. Holmberget al., “TidyBot++: An open-source holonomic mobile manipulator for robot learning,” inConf. on Robot Learning (CoRL), 2024

    work page 2024

  25. [25]

    TacFR-Gripper: A reconfigurable Fin Ray-based com- pliant robotic gripper with tactile skin for in-hand manipulation,

    Z. Xuet al., “TacFR-Gripper: A reconfigurable Fin Ray-based com- pliant robotic gripper with tactile skin for in-hand manipulation,” arXiv:2311.10611, 2023

    work page arXiv 2023

  26. [26]

    Learning fine-grained bimanual manipulation with low-cost hardware,

    T. Z. Zhao, V . Kumar, S. Levine, and C. Finn, “Learning fine-grained bimanual manipulation with low-cost hardware,” inRobotics: Science and Systems (RSS), 2023

    work page 2023