A reality-capture model of your bridge or dam, streamed from a dedicated cloud GPU to Chrome or Edge. Navigate it, measure it, review AI-proposed defects, sign off findings.
A capture of a bridge or a dam runs to hundreds of gigabytes. AerTwin keeps it on the server and sends a 3D interactive video of it instead.
The model renders on a cloud GPU while an inspector is in it, and is released when they leave. A current browser is all your team needs.
The browser receives video, screen coordinates and record numbers. No mesh or point cloud is ever transmitted.
Measurements are computed on the server, against the model itself. The same figure appears on screen, in the PDF and in the CSV.
Several people can be in the same structure at once. A saved finding syncs across, a bell says what changed, and the activity panel holds 30 days — click an entry to jump straight to that finding.
The tools do not change with the structure. Findings are filed under its own parts, so the report reads the way your inspection does.
Move along the deck and down each pier, and file findings under Deck, Pier 2 or Bearings — those groups become the sections of the report.
Work across the whole face at once, and file findings under Upstream face, Spillway or Crest — the same way a bridge is filed by its parts.
Three steps, and only the last one reaches your computer.
Photographs and LiDAR are captured together and reconstructed to scale into one textured 3D model.
A cloud GPU loads the model and renders exactly the view the inspector is looking at, at the size of their screen.
The view arrives as encrypted video. A measurement is computed on the server and returned as numbers.
Orbit, pan and zoom with CAD-style controls, or double-click any point to fly to it. Every finding keeps the viewpoint it was recorded from, so anyone can return to what the inspector saw.
Distance, area, volume, depth profile and 3D defect tools trace against the model on the server — not against a picture of it.
Our computer vision sweeps every survey photograph before your team opens the model. Each proposal is placed on the 3D model, measured, and linked to the photographs it came from.
Each finding carries its component, severity, status, recommended action and target date. File them under the parts they belong to — Deck, Pier 2, Spillway — and those groups become the sections of the PDF.
Pin each piezometer, crackmeter, water-level gauge or accelerograph where it actually sits on the structure. Every type has its own symbol, so an engineer can read a structure they have never visited, and one click on the list flies to any instrument.
Three account types — Administrator, Inspection Engineer and Viewer — so a critical finding raised by one engineer is reviewed by another, and the record shows who reviewed what.
AerTwin shows the structure. The Planner is the part beside it that turns what you found into a work order, follows it to site, and keeps the evidence attached to the place it belongs to — the photographs your team took, and the model they were taken against.
Preview of the next AerTwin release. Ask us what is running today and what is still being built — we will show you both.
Managed cloud servers, with a session placed on the one closest to your team. Nothing to install, patch or lose.
The model, its textures and the measurement engine stay in the data centre. The browser receives encrypted video and numbers.
Every person signs in with their own account, a password and an authenticator app. It cannot be switched off.
Permissions are read at the moment of each action. Changing someone's role ends their live session.
GPU machines accept no inbound web connections and never receive a user's login token.
A 3D model cannot resolve it at any zoom, so AerTwin opens the original photograph instead.
Zooming magnifies the picture. The sampling resolution is printed next to every volume.
Past 60% uncertainty a volume reads Not measurable. A refusal is a correct answer, not an error.
A detector's miss rate is not known. Every report states how many proposals are not yet reviewed.
Edit a signed-off finding and the sign-off is withdrawn until someone accepts the new one.
What is verified, how it was checked, and what is not. Ask for it before you rely on a figure.
Request the accuracy statement →Bridges, dams and comparable civil structures captured by drone. The tools are the same on each — only the model changes.
No. AerTwin gives you the evidence behind a rating — each finding located, measured, photographed, with a severity and a named sign-off — and exports it as PDF and CSV for your own process.
A current version of Chrome or Edge, a steady broadband connection, and UDP allowed through your firewall. A dedicated graphics card is not needed. We test your network with you before any pilot.
Your session runs on the cloud server closest to your team — today that is our European data centre.
Very little. We study the structure from what you already hold — drawings, past reports, earlier surveys — together with what is publicly available, then carry out the data collection ourselves, build the model from it and host it on our servers.
We will stream a live model to your screen, measure it together, review AI-proposed defects and export a report — including the parts that say what the model cannot measure.
Or write to info@aer-intel.com