EN
中文
EN
中文
专为人形机器人设计的通用底层架构，实现软硬件深度解耦，支持跨平台快速部署。提供统一的API接口与标准化的数据格式，确保应用层逻辑在不同机型间的无缝迁移。 A general underlying architecture exclusively designed for humanoid robots, enabling deep decoupling of hardware and software and supporting rapid cross-platform deployment. It delivers unified APIs and standardized data formats to ensure seamless migration of application-layer logic across different robot models.
上传动捕数据、舞蹈视频，或使用一句文字、一段音乐，生成机器人可执行的动作序列。 Upload motion capture data, dance videos, or use a sentence of text or a piece of music to generate a sequence of actions executable by a robot.
基于视觉识别与实时规划，机器人自主完成物体检测和稳定抓取。 Based on visual recognition and real-time planning, the robot autonomously performs object detection and stable grasping.
拖拽机器人关节，精细复刻示教动作。 Drag the robot joints to precisely replicate the teaching motions.
导入室内地图，标记起点和途经点，快速规划巡检或导览路线。 Import an indoor map, mark the starting point and waypoints, and quickly plan inspection or tour routes.
利用机器人的激光雷达，实时扫描环境构建地图。 Use the robot's LiDAR to scan the environment in real time and construct a map.
通过机器人相机，识别人物或手势。 Use the robot's camera to recognize people or gestures.
内置语音智能体，实现自然语言对话与实时指令响应。 A built-in voice agent enables natural language dialogue and real-time command response.
通过视觉感知，机器人识别目标人物并保持安全距离跟随。 Through visual perception, the robot identifies the target person and follows while maintaining a safe distance.
ODance Studio 运动控制工具平台，集关键帧编辑、一键轨迹插值、腰足姿态精细调校、动作仿真推演功能于一体，可高效完成机器人各类动作编排调试；同时内置运动风险检测与轨迹分析能力，能够提前排查碰撞、失衡等隐患，量化运动数据，大幅提升人形机器人动作开发效率与整机运行安全稳定性。
ODance Studio supports keyframe editing, one-click trajectory interpolation, waist-foot fine-tuning and motion simulation for fast robot motion debugging. Its embedded risk detection and trajectory analysis pre-empt collisions and instability, quantifies motion data, and boosts humanoid robot development efficiency and operational safety.
自研通用底层软件与统一通信协议，实现软硬件解耦，支持不同品牌与型号机器人的极速适配与高效部署。
Self-developed general-purpose underlying software and unified communication protocols enable hardware-software decoupling, supporting rapid adaptation and efficient deployment of robots of different brands and models.
基于模仿学习和强化学习的全身控制，实现高难度、高协调性的复杂动作生成。
Based on imitation learning and reinforcement learning, whole-body control enables the generation of complex motions that are highly difficult and well-coordinated.
基于几何感知、深度学习与视觉语言模型，实现对物理世界的深刻理解。
Based on geometric perception, deep learning, and vision-language models, achieve a profound understanding of the physical world.
融合视觉、雷达的感知输入与多模态模型的推理能力，实现“听得懂指令”的自主环境探索与导航。
By integrating visual and radar perceptual inputs with the reasoning capabilities of multimodal models, the robot achieves instruction-aware autonomous environment exploration and navigation.
基于大模型调度原子Skill，实现对复杂或模糊指令的意图拆解与任务执行。
Leveraging large models to orchestrate atomic skills, it enables intent decomposition and task execution for complex or ambiguous instructions.
基于基础能力，搭建面向教育、陪伴、接待、物流等多场景应用。
Based on foundational capabilities, build applications for multiple scenarios such as education, companionship, reception, logistics, and more.
橡树清溪科技 OakCreek Tech 成立于2025年6月。
OakCreek Tech was founded in June 2025.
致力于打造人形机器人通用系统及应用平台，全套即插即用能力覆盖底层操作系统、全身运动控制、跨环境语义导航、长程任务意图理解、多模态高阶决策，及多场景应用等关键环节。
Our company is committed to developing a general-purpose system and application platform for humanoid robots. Its complete plug-and-play capabilities cover core modules including operating systems, whole-body control, semantic navigation, long-term task intention comprehension, high-level multimodal decision-making, and multi-scenario applications.
核心团队汇聚了来自Stanford、MIT、UC Berkeley、清华、北大等顶尖学府的学术精英，以及微软、阿里、Momenta等一线科技企业的资深专家，兼具前沿科研视野与丰富产业经验。
Our core team consists of academic elites graduated from top institutions including Stanford, MIT, UC Berkeley, Tsinghua University and Peking University, as well as senior specialists from leading tech enterprises such as Microsoft, Alibaba and Momenta, boasting cutting-edge research insights and extensive industrial experience.
橡树清溪OakCreekTech
OakCreek Tech
橡树清溪OakCreekTech
OakCreek Tech
橡树清溪OakCreekTech
OakCreek Tech
地址：北京市海淀区清华科技园创业大厦207
Address: Room 207, Valley Mansion, Tsinghua Science Park, Haidian District, Beijing
电话：19072974147
Phone:+86-19072974147
邮箱：support@oakcreektech.com
Mail：support@oakcreektech.com
