sizeless

sizeless is an AI-powered spatial tool for underground infrastructure that turns smartphone videos into 3D models and CAD plans.

Sizeless is a spatial AI workflow designed for underground construction, district heating, and utility connections. It creates precise 3D digital twins of open trenches and house connections directly for GIS and CAD systems. Using an iPhone Pro, workers capture video footage at the construction site without needing separate surveying appointments. Developed with algorithms from ETH Zurich, the system processes these videos into high-resolution point clouds, DWG and DXF revision plans, and 3D models. The platform enables autonomous documentation by installation crews, eliminates manual sketches, and ensures GDPR compliance with data storage in the EU.

sizeless at a glance

Best for: Researcher, Software engineer
Runs on: iOS
Pricing: Freemium
Released: September 2026
Category: AI research tool
Website: sizeless.co

Key features

Smartphone video capture for trenches and utility connections
High-resolution 3D point cloud generation
Norm-compliant 2D CAD revision plans in DWG and DXF formats
3D models and GIS integration for pipe routing
GPS-independent operation in cellar areas and basements
Automated measurement of earth masses and dimensions
GDPR-compliant data storage in the EU

Use it for

Documenting open trenches and house connections for GIS and CAD
Accelerating construction time and billing processes
Creating reliable 3D evidence including third-party utilities
Detecting construction errors before trench backfilling

Alternatives to sizeless

Nomic Atlas — It is purpose-built to parse and reason over complex drawings, specs, and CAD files.
SAG — It combines model routing, knowledge management, and agent orchestration in private.
Amagine3D — Generates editable CAD source code directly from natural language and images.
Geo Sleuth — It uses geometry, elevation data, and open street maps to locate photos without metadata.
UltraRAG — Build complex RAG workflows using modular MCP components and YAML configuration.
Autonomyware — Engineers physical products autonomously from ideas and sketches.
AutoIllustrate — Produces fully editable vector diagrams and enhanced figures for research papers.
Parseable — An open, cost-effective observability platform with AI-native querying.
Project NOMAD — Provides access to large-scale knowledge bases without requiring an internet connection.
AntV Chart MCP — Provides direct integration of diverse chart generation tools into AI applications.
Cube — It keeps metric definitions consistent across AI agents and human users.
DeepAnalyze — Automates the entire data science pipeline including analysis and visualization.

All sizeless alternatives