spatial-ai

🐍 Geometric Computer Vision Library for Spatial AI

InternRobotics' open platform for building generalized navigation foundation models.

Official implement of VGGT-Long

Generalizable Perception Stack for all things 3D, 4D & Scene Understanding

Open Reality: phone video to AI-queryable 3D scenes. MCP tools for Claude/Codex/Cursor (npm openreality-mcp), the self-hostable broker, and the VGGT-SLAM library, in one repo. BSD-2-Clause.

Terra: Hierarchical Terrain-Aware 3D Scene Graph for Task-Agnostic Outdoor Mapping

Real-Time Spatio-Semantic Memory

MLX-native 3D and spatial inference for Apple silicon: object reconstruction, image-to-mesh, scene geometry, Gaussian splats, and multi-view bundles.

A Synthetic Benchmark for Evaluating Spatial Intelligence in Visual Foundation Models

Spatial analysis using Deep Spatio Temporal Point Process (DeepSTPP) updated with state-of-the-art mamba state space model, using spatiotemporal Hawkes Process, spatiotemporal self-correcting process, NJ COVID-19 and Japanese earthquake data.

Feed-forward video-to-3D scene reconstruction using VGGT (CVPR 2025) with open-vocabulary semantic labeling via SAM 2.1 + CLIP. One command, no COLMAP, Apple Silicon + CUDA.

3D scene reconstruction from monocular video — COLMAP + 3D Gaussian Splatting + Grounded-SAM-2, with open-vocabulary CLIP queries and an agentic QA release gate.

Real-time 3D Hand & Body Pose Estimation with NVIDIA FoundationStereo & RTMPose for Dual-Eye Intelligence Perception (DIP)

Stochastic Preconditiong for Neural Field Optimization -- additional website assets

A curated benchmark zoo for embodied navigation tasks, datasets, metrics, leaderboards, and reproducibility notes.

repo to describe the floor layout project

SpAItial — independent third-party profile of a public API surface, by API Evangelist. SpAItial is a frontier AI lab building physically-grounded world models. Its Echo model turns a single image, a 360-degree panorama, or a text prompt into a persistent, explorable 3D Gaussian Splat world. The SpAItial Developer API is a REST interface for generat

Autonomous agent for SF design-build firms: finds AI-homeowner leads, generates 3D remodel proposals, drafts outreach, and learns from outcomes.

Reproduce Neural Assets in PyTorch with a clean, paper-faithful implementation for MoVi training and research

AI-powered indoor spatial navigation — Where Google Maps ends, Inevi begins. Live camera location detection, conversational AI guide, multilingual (EN/TE/HI), built with Groq, AWS Aurora DSQL, DynamoDB, S3, and Vercel.