Motion generation tools produce candidate trajectories, but interactive robot authoring also requires recoverable execution, traceable edits, explicit geometry, and model-specific validation. We present Kinora, a system that connects text-to-motion generation, a motion library, local agent tools, robot retargeting, and trajectory refinement through versioned artifacts. The current release uses an independently specified 52-joint authoring rig and procedural skeleton previews, separating the distributed human geometry from externally installed generation models and robot assets. Human-motion edits preserve frames outside a selected interval, while robot edits create immutable candidates whose reports govern acceptance and export. We evaluate server 0.5.1 and Gadget 0.15.2 with 156 core regression test executions, 49 Gadget executions, four CUDA generation/editing cases, a 30-frame G1 retargeting check, and a 3,600-frame synthetic G1 fixture paired with a suspended negative control. All four CUDA cases passed their numerical gates; the turning splice nevertheless retained a 14.3 cm support-foot drift proxy warning. The G1 positive fixture completed bounded simulation replay, whereas the suspended control was quarantined before replay. These results support selected integration and preservation contracts. They do not establish semantic-quality improvements, population-level feasibility, or hardware safety.
Figures & tables
Figure 1: Kinora’s artifact workflow. Validation belongs to a particular candidate and robot configuration. A failed or unavailable check does not become a successful result through export or a different client interface.
Check
Result
Scope
Core regression
156 passed
Backend protocol, edits, duration, robot editor/dynamics, account and permission paths.
Gadget regression
49 passed
Local protocol, HTTP permission behavior, and recovery.
Pinned external assets
29 verified
G1 robot files only; content hashes checked.
Independent human rig
7 tests passed
Offline regeneration, FK consistency, recovery, rejection, pivot preservation, export, and splice checks; included in the core count.
G1 retargeting
30 frames
Actual generated-source conversion and finite robot output; not a dynamics certificate.
CUDA diagnostic cases
4 passed
Numerical gates and selected edit invariants; quality warnings retained.
Table 1: Release checks and their evidential scope. Counts are test executions or inspected assets, not independent samples of real-world reliability.
Case
Step (m)
Accel. (m/s 2 )
Rotation (rad)
Drift (cm)
Exact cut
Repair
0.01023
2.086
0.0364
—
Yes
Splice, straight
0.07118
63.888
0.3568
3.1
No
Splice, turn
0.01771
15.939
0.0653
14.3
No
Seam rejection bound
0.20000
120.000
0.5236
—
—
Table 2: Measured edit-seam diagnostics. Maxima aggregate the start and end seams within each case. All three edits preserve the exterior exactly. “Exact cut” refers to the recorded preservation-through-acceleration flag, not the broad numerical pass condition.
Quantity
Positive fixture
Suspended control
Limit
Root-force ratio
2.15×10−5
1.00001
0.01
Root-torque ratio
8.93×10−5
0.02052
0.01
Inverse-dynamics actuator utilization
0.14306
0.04843
1.0
Joint-speed utilization
0.00102
0.00032
1.0
Replay actuator utilization
0.16023
—
1.0
Replay joint-position error (rad)
0.00137
—
0.05
Table 3: G1 validation results. Replay quantities for the suspended control are unmeasured because prerequisite dynamics checks rejected it. Lower values indicate smaller residuals, utilization, or tracking error; they are not cross-model quality scores.
Robot motion generation creates feasible, task-directed trajectories for robots to move, manipulate objects, and interact with people and environments.
Language-conditioned humanoid motion generation enables users to create and refine robot movements through natural language. The central challenge is to connect an underspecified request to concrete motion structure that can subsequently be revised relative to an accepted reference. We present TRACE, a generation and editing framework for the Unitree G1 humanoid robot built around typed, measurable kinematic quantities. Our insight is to ground generation conditions in paired training motions and editing goals in the current motion. A language model learns to predict support, root, and hand events from motion-derived supervision; their time-aligned encodings complement the original text in a diffusion generator. Gravity-referenced states and trajectory--contact objectives improve kinematic consistency. For editing, typed requests define body--time scopes and source-relative goals. A separate reference-conditioned diffusion editor uses these goals and revised events to realize changes, while sampling-time projection preserves protected state coordinates. Quantitative evaluation and human assessment show improved semantic alignment and motion quality over the compared methods. Multi-turn editing evaluations demonstrate more reliable execution of successive relative requests under shared preservation constraints. Code is available at https://github.com/Multimodal-Perception-Action/TRACE.
Zejian Kang, Yifu Luo, Boyang Shen +9
Zhejiang University · Westlake University · Technical University of Munich +7
General-purpose humanoid trackers can execute diverse references, but their zero-shot coverage depends on large embodied corpora that are costly to extend. Text-to-motion generators offer scalable supervision, yet models trained on human motion or retargeted data inherit a gap between kinematic plausibility and robot executability. Existing one-way pipelines fix either the generated corpus or the reward tracker. We introduce GenTrack, an online generator--tracker framework that alternates execution-grounded, group-relative generator alignment with tracker training on newly generated references; anchoring and rehearsal constrain drift. On Unitree G1, we evaluate GenTrack with ProtoMotions and SONIC backbones across three zero-shot tracking splits including public AMASS and LAFAN benchmarks, and a private out-of-distribution test set of 1,024 prompt-motion pairs in the wild. The online co-training strategy consistently produces generators that output more robot-executable motions with strong semantic alignment, and trackers with markedly broader zero-shot coverage and improved tracking accuracy, especially on out-of-distribution references. These results demonstrate that joint online post-training effectively narrows the executability gap between retargeted references and robot-native motion, advancing zero-shot humanoid control without additional data collection and beyond the limitations of a static reference pool.
Zeyu Ling, Xinyao Yu, Renye Yan +4
Zhejiang University · Peking University · Tencent +1
Text-conditioned motion generation is a promising interface for programming humanoid robots, yet current generators are often trained on human motion datasets retargeted to robot morphologies. Although such data provides rich semantic and kinematic priors, it fails to capture the nuances of whole-body tracking controllers, including balance, contact dynamics, actuation limits, and controller-specific failure modes. As a result, generated motions can be semantically plausible but difficult or impossible for the robot to execute. We introduce TEXEDO, a test-time scaling framework for humanoid motion generation that improves motion quality without requiring a stronger underlying generator. Given a text prompt, TEXEDO samples multiple candidate motions from a pretrained text-conditioned generator and selects the best motion that is both executable and task-aligned. The reward model combines a dynamic feasibility verifier, distilled from whole-body tracking rollouts to predict physical executability, with a semantic alignment verifier that measures text-motion alignment in a learned co-embedding space. Our pipeline treats dynamic feasibility as a hard constraint and semantic alignment as the selection objective within the feasible set. Through large-scale simulation studies and real-world deployment on a Unitree G1 humanoid robot, we show that TEXEDO consistently improves both tracking fidelity and text alignment. These results demonstrate that grounded verification is an effective path toward deployable language-guided humanoid motion generation. Project website: https://jianuocao.github.io/TEXEDO/
Jianuo Cao, Yuxin Chen, Yuzhen Song +3
Nanjing University · University of California, Berkeley · Southern University of Science and Technology