# SpatialLM / 空间智能大模型 — AI Compare Ninja

> 3D language model from Manycore Tech that turns point clouds into structured indoor layouts with walls, doors, windows and objects.

## Overview
- **Category:** Local AI (Apps & interfaces)
- **Status:** Active · checked
- **Paid from:** Not verified for this market
- **Free access:** Free, non-commercial use only (CC BY-NC 4.0) (The SpatialLM1.1 weights are under CC BY-NC 4.0 and may be used only for non-commercial purposes (Hugging Face model card and GitHub readme).)
- **Website:** https://manycore-research.github.io/SpatialLM/
- **Canonical URL:** https://www.aicompare.ninja/en/tool/spatiallm

## What it does
- Detects walls, doors and windows plus oriented object boxes with semantic categories in 3D point clouds.
- Handles point clouds from monocular video, RGB-D images and LiDAR sensors; for an RGB video, MASt3R-SLAM builds the point cloud.
- Outputs 3D bounding boxes for walls, doors, windows and objects; the project page names 2D floor plans and IFC only as a possible extension (code for this not published). Weights and code are on GitHub and Hugging Face.

## In brief
3D language model from Manycore Tech that turns point clouds into structured indoor layouts with walls, doors, windows and objects. Detects walls, doors and windows plus oriented object boxes with semantic categories in 3D point clouds. Handles point clouds from monocular video, RGB-D images and LiDAR sensors; for an RGB video, MASt3R-SLAM builds the point cloud. The SpatialLM1.1 weights are under CC BY-NC 4.0 and may be used only for non-commercial purposes (Hugging Face model card and GitHub readme).


## Limits
Weights and code are available on GitHub and Hugging Face and run on the user's own hardware. The SpatialLM1.1 weights are under CC BY-NC 4.0 and may be used only for non-commercial purposes (Hugging Face model card and GitHub readme).

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Open weights (CC BY-NC 4.0) | free | Permanent, non-commercial only | SpatialLM weights and code free on GitHub and Hugging Face; CC BY-NC 4.0 licence: no commercial use |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://manycore-research.github.io/SpatialLM/
- https://github.com/manycore-research/SpatialLM
- https://huggingface.co/manycore-research/SpatialLM1.1-Qwen-0.5B
- https://manycoretech.com/

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/spatiallm*
