AI agents that actually understand engineering drawings.
Automate design checks, process calculations, bill-of-material validations, design-for-manufacturing analysis, and other engineering workflows across drawings, CAD files, and datasheets.
AI agents are rewriting how software gets built. The same shift can't happen in the physical world until the documents that contain engineering knowledge become machine-readable and designs becomes verfiable.
Trusted by Fortune 500 companies in manufacturing and engineering.
Drawing and Document Analysis
to view metadata
Our Services
Extract structured data from engineering documents
Our parsing API converts P&IDs, spec sheets, BOMs, datasheets, and technical drawings into structured JSON with higher accuracy than general-purpose vision-language models. Use our API, SDK, or human-assisted conversion services.
Intuitive interfaces for data validation
AI-assisted user interfaces to check and fix connectivity, classification, numerical values, and more for ingested P&IDs, single line diagrams, BOMs, and other engineering documents.
Build AI agents for engineering workflows
Our agent builder comes with prebuilt tool calls for graph analysis, flow analysis, Monte Carlo simulations, numerical calculations, and more. Easily extensible for DFM analysis, process flow simulations, engineering calculations, and custom workflows.
Built-in and customizable ontologies
Pre-built but extensible schemas for BOM trees, supply chains, EPC sites, facilities, and more. Map your engineering data to structured ontologies that fit your domain without starting from scratch.
Who We Are
Our founding team has worked on foundation models for 3D CAD, manufacturing execution systems (MES), all sorts of robot hardware manufacturing, defense systems, disaster risk modeling, and neural fields (used for surrogate models, 3D reconstruction and more)... at places like NVIDIA, Meta, Oracle, Embark Trucks, Tulip Interfaces, and more.
Towaki Takikawa
Co-Founder / CEO
I've been building machines since I was a child. First robots, then autonomous vehicles, then artificial intelligence that can understand the 3D world around us. I worked as a research scientist at NVIDIA, leading discoveries in 3D foundation models and spatial understanding at. Our work has been cited over 4,500 times across papers published at CVPR, ICCV, SIGGRAPH. I also worked on Kaolin, NVIDIA's open-source 3D machine learning library. I left because I wanted to work directly on what I felt most passionate about... unlocking creative freedom that is blocked by the repetitive work that goes into design verification.
Book a demo.
Trusted by manufacturing and engineering teams to extract structured data from legacy documents and build AI agents.
Contact us at: info@outerport.com