nvidia-isaac/cuVSLAM

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

1773 stars · 196 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

cuVSLAM is NVIDIA's CUDA-accelerated library for real-time visual odometry and simultaneous localization and mapping (SLAM). It supports multiple tracking modes including mono, RGB-D, multicamera, stereo-inertial, and multisensor configurations, with Python, C++, and ROS2 APIs.

Use cases

• real-time visual odometry for robots
• SLAM for autonomous navigation
• stereo camera tracking
• RGB-D camera odometry
• visual-inertial odometry with IMU fusion
• multi-camera tracking rigs
• robotics localization and mapping

When to choose

• you need GPU-accelerated real-time SLAM on NVIDIA hardware
• you want a library that works out of the box without parameter tuning
• you need support for diverse sensor configurations (mono, stereo, RGB-D, IMU)
• you're building robotics applications within the NVIDIA Isaac ecosystem

When to avoid

• you don't have NVIDIA CUDA-capable GPU hardware
• you need CPU-only SLAM solutions
• you require a permissively licensed library (license is custom/NVIDIA)
• you need a fully open-source modifiable implementation

Facets

library · maturity active

computer-vision machine-learning simulation graphics robotics computer-vision autonomous-vehicles gpu-computing cpp python slam visual-odometry cuda mapping vio stereo-vision nvidia-isaac real-time linux gpu ros

Member repositories

For agents

markdown · JSON · MCP: product_card(name="nvidia-isaac/cuVSLAM")