An agent that edits and queries 3D spatial data - so you spend less time scripting pipelines and more time thinking about the data itself.
Ninety seconds, no narration: the agent loads a scan, runs your edits, answers a spatial question, and exports - all in plain English.
Working with point clouds today means stitching together CloudCompare, PDAL, Python, and a stack of half-remembered shell commands.
panoriq replaces that loop with a single agent you describe your intent to. It plans the operations, runs them, shows you the result, and lets you iterate until the data looks the way it should.
Edit, query, and trust - the three operations that turn a raw point cloud into something you can act on.
"Remove the trees on the north side." "Downsample to 1 cm in this region." Describe the change, get the result.
"How tall is this structure?" "How many objects are in this scene?" Get answers grounded in the underlying geometry, not approximated from a screenshot.
Every action the agent takes is recorded as a structured operation - re-runnable, auditable, diffable. Eval-driven, not vibes-driven.
Load your data, describe what you want, iterate until it's right. The agent handles the operations - you stay in the loop.
Point the agent at your data - local files or a remote scan. Nothing leaves your environment unless you explicitly send it somewhere.
Describe what you want in plain English. The agent breaks the request into operations, runs them, and shows a diff of what changed before you commit.
Refine, branch, undo. When the result is right, export to your format of choice or pipe it straight into the tools you already use.
"The right interface turns a lidar scan into a conversation - and a step toward systems that operate in physical space."