Quickly Build World-Class Motion Capabilities for Your Robot—RoboCraft AI Beta Now Open for Applications
Robot model adaptation · motion customization · rapid deployment


We have been thinking: can robot motion development become as simple as editing videos or documents? Facing humanoid robots of different brands and configurations, can the industry have a unified, complete development toolchain and deployment process, and build continuously evolving motion capabilities for robots?
The reality is that both hardware manufacturers and purchasers face different industry challenges.
For hardware manufacturers,developing motion control algorithms from scratch is a costly uphill battle. Under the traditional model, motion control development is an extremely demanding"multi-disciplinary capability and engineering experience"black hole, requiring teams to master dynamics control, reinforcement learning, simulation modeling, and Sim2Real deployment. And the implementation of motions relies heavily on the hands-on experience and troubleshooting abilities of expert teams. This is a asset-heavy, long-cycle endeavor, requiring significant human and time investment.
For downstream purchasers in the industry chain, due to the closed and fragmented upstream ecosystem, they need secondary development capabilities to meet different user demands in real-world scenarios. But in the past, this was highly experience-dependent, requiring engineering and algorithmic skills, which was not friendly to most robot users.
At this stage, the industry needs not just breakthroughs in individual algorithms, but a unified, reusable, standardized motion development platform.
RoboCraft AI:
A General-Purpose Robot Motion Capability Development Platform
Based on the above industry pain points, BridgeDP introduces RoboCraft AI—a general-purpose robot motion capability development platform.
It distills BridgeDP's algorithm accumulation and Sim2Real engineering experience from over 50 robot projects across different brands and configurations, transforming expert-level technology into unified, reusable platform capabilities and standardized development processes.
We hope developers can use this platform to quickly build world-class motion capabilities for robots.
Here is a demonstration of RoboCraft AI's development capabilities:
Addressing the dual pain points of hardware manufacturers and downstream application developers, RoboCraft AI solves problems from two dimensions.
Path One:
0→1 Motion Control Integration for Self-Developed Robots
For enterprises developing their own robot hardware, RoboCraft AI provides a fast, unified integration process, gradually standardizing steps that previously relied heavily on expert experience:
- Intelligent validation of URDF modeling rationality: Robot structural issues often only surface when testing on physical robots, posing safety risks. The platform can identify, flag, and fix problems at the URDF stage in advance.
- Hardware safety procedures and inspection mechanisms: Non-standard hardware issues are common. BridgeDP has accumulated debugging experience across multiple robots and transformed it into standardized safety inspection procedures.
- Automated hyperparameter tuning during training: Motion control model training requires repeated parameter tuning for different motions and environments, which is extremely costly. The platform automates parameter tuning, significantly reducing manual intervention.
After integration, self-developed robots gain basic motion control capabilities and a development environment consistent with mature models, significantly shortening the cycle from hardware availability to sustainable development.

Beta support scope: applications are open for self-developed robot integration, with specific robot models and timelines subject to review and evaluation.
Path Two:
Motion Choreography and Application Deployment for Mature Robot Models
For users who have purchased mature robot models (rental companies, integrators, content creators, research teams, etc.), RoboCraft AI provides an out-of-the-box motion development toolchain, eliminating the need for repeated integration and training processes.
Motion acquisition and choreography:
- Supports uploading MP4 videos to extract motions and importing BVH motion files;
- Visual timeline choreography with low-code motion combination and commercial performance flow configuration;
- Built-in motion asset library management, supporting team-wide accumulation and reuse.
Multi-robot deployment and swarm control:
- Batch robot deployment, one-click application distribution
- Multi-robot coordinated startup and standby
Team collaboration and asset management:
- Unified management and sharing of motion assets, application templates, and robot configurations;
- Assets developed by members can be added to the library with one click, avoiding redundant development.

Beta compatible robot models: Unitree G1, AGIBOT X2; more brands and models are being added continuously.
Beta Applications Are Now Open
During the beta period, we will provide the following benefits:
- Free access to and deployment of multiple popular dance applications for models such as Unitree G1 and AGIBOT X2;
- Unlimited use of custom motion features.
We especially welcome the following groups to apply for the beta:
- Robot hardware R&D companies looking to quickly gain mature motion capabilities;
- Robot rental companies and system integrators looking to quickly deploy robot applications for commercial performances, cultural tourism, and educational demonstrations;
- Universities and research institutions looking to conduct robot research and teaching;
- Next-generation developers looking to explore robot content creation, motion development, and application innovation.
Scan the QR code to apply:

After approval, users can register, log in, and create organizations on their own. We will also invite users to join the RoboCraft AI core beta community to work with the BridgeDP team to drive product iteration.





界面新闻
深视新闻
深圳特区报
财新
36氪
钛媒体
中国新闻周刊
蓝鲸新闻
Z Finance
投资家网
雷峰网
亿欧
环球网
南方日报
极客公园



