Two quality levels
Standard builds a fast, locally consistent map for most use cases. Precision runs the full GLIM export path (map.ply → output.pcd) with trajectory output for cases where accuracy matters most.
LiDAR SLAM
GLIM-quality LiDAR mapping, without the ops burden.
LidarFlow runs GLIM SLAM — one of the most respected open-source LiDAR SLAM engines, from Kenji Koide at AIST — on your recordings in the browser. Standard for fast iteration, Precision for full-fidelity exports.
Works without GPS
Locally consistent map even without GNSS. Add a GPS topic and the same run can be georeferenced to WGS84.
Intensity preserved
If your LiDAR captured intensity, it survives into the output point cloud. A guarantee, not a best-effort.
Field-tested presets
Ouster, Livox, and Velodyne-class sensors are covered out of the box. The validator detects what you uploaded and suggests the right preset.
GLIM · LiDAR SLAM
What you upload
- · `.mcap` or ROS1 `.bag`, up to 20 GB per file
- · LiDAR PointCloud2 topic (required)
- · IMU and GNSS topics (optional, auto-detected)
What you get back
- · `output.pcd` — mapped point cloud
- · `map.ply` — full-fidelity export (Precision runs)
- · `slam_trajectory.txt` — sensor path through space
- · `metadata.json` — settings, validation, provenance
Same GLIM, none of the ops
Upload a LiDAR rosbag. Get a mapped point cloud back.