Wednesday, November 11, 2026
Product news

“Inspect this building”: AI tool turns simple instructions into complete drone flight plans

Planning a roof inspection, creating a 3D model of a building or mapping an area can be time-consuming. Skyplanr aims to simplify the process with AI-powered mission planning. Developed by Norwegian drone pilot and software developer Sander Helgesen, the tool allows users to describe a task in plain language. Skyplanr then translates that instruction into a complete drone flight plan, including preflight checks.

From instructions to automated flight plans

The concept behind Skyplanr is straightforward. Users select a location on the map, choose the drone they intend to fly and describe the task they want to perform. This can be done using predefined mission types, such as real estate photography, inspections or mapping, but users can also simply enter a natural-language instruction.

For example, typing “Inspect this building” or “Create a 3D model of this building” is enough to get started. An AI agent then translates the request into a flight plan containing waypoints, flight altitudes, camera angles, speeds and photography settings.

The software takes several environmental factors into account, including terrain elevation, building heights, tree heights, airspace restrictions and the position of the sun. Weather conditions are also considered when planning a mission.

The software actually thinks along

One of Skyplanr’s more interesting features is that it does more than simply generate a series of waypoints. It also attempts to account for practical conditions at the selected location.

During one of my test missions, the software automatically suggested an alternative take-off location because trees were located just a few metres from the originally selected spot.

The explanation provided alongside the flight plan was equally detailed. For a 3D mapping mission of a sports hall, Skyplanr proposed a combination of a straight-down grid and a second grid with an oblique camera angle. The plan consisted of 358 photographs in total, divided into multiple battery flights. The software also calculated the required flight altitude to achieve a ground sampling distance (GSD) of approximately 2 cm per pixel.

Importantly, Skyplanr explains why it makes certain decisions, including the reasoning behind the selected take-off location, flight altitude and camera settings.

106 automated checks

Once the AI agent has generated a flight plan, it is passed through a separate validation system. According to the developer, each mission is subjected to 106 automated checks.

These checks cover aspects such as obstacle clearance, flight altitude, distance from the take-off location, battery capacity and Return-to-Home altitude. The system also considers weather conditions, airspace restrictions and potential issues affecting people on the ground. The flight path is checked at intervals of just two metres, rather than only at individual waypoints.

The AI therefore does not simply decide whether a mission is safe to fly. A separate validation system checks the proposed route against predefined criteria. Nevertheless, the pilot remains responsible for reviewing the flight plan, checking current conditions and ensuring that the operation complies with applicable regulations.

First impressions: surprisingly easy to use

I tested Skyplanr myself by creating a mapping mission for a nearby building, using a DJI Mini 5 Pro as the selected aircraft. Creating the mission was remarkably straightforward: within just a few clicks, a complete flight plan was ready.

The ability to describe a task in plain language is particularly appealing. Instead of manually placing waypoints and configuring individual parameters, you simply tell the software what you want to accomplish.

Modifying an existing flight plan is also straightforward. Users can adjust settings, move waypoints or make other changes without spending additional credits. Only the generation of a new AI mission counts towards the monthly allowance.

One aspect that is currently less straightforward is transferring the generated mission to the drone controller. Depending on the controller and operating system, this may still require several manual steps.

A dedicated iPhone app that installs missions directly into DJI Fly is currently available in beta. For certain DJI controllers, a manual transfer is still necessary.

Developed by a drone pilot

Helgesen combines his background in software development with his work as a drone pilot. He develops apps independently and primarily uses a DJI Air 3S for real estate photography.

AI played a major role in building Skyplanr. Helgesen started development on September 30, using Anthropic’s Claude Code as his main development tool. According to him, the first version was ready after approximately three weeks and around 1,000 commits.

The AI handled most of the programming, testing and documentation. Helgesen himself decided which features to develop and tested the software in real-world flying conditions. His practical experience was particularly important when dealing with aviation regulations, the intricacies of DJI’s KMZ format and the way flight plans behave on actual controllers.

It is an interesting example of how AI-assisted development can accelerate the creation of specialised software. Skyplanr combines a web interface, an AI-powered mission planner, a validation system and support for several drone platforms.

Mapping and 3D processing

Skyplanr is not limited to flight planning. After completing a mission, users can upload their photographs for processing. The platform can generate orthophotos, elevation models, point clouds and textured 3D models. This functionality is based on OpenDroneMap and is currently available free of charge during the beta phase.

There are, however, some geographical limitations. The accuracy and availability of elevation data vary between countries. Skyplanr combines different open datasets and elevation models to estimate terrain, building and tree heights. In some regions, detailed LiDAR data is available, while in others the software has to rely on less precise sources. As a result, not all obstacles may be represented accurately. Power lines, recently constructed buildings and other changes to the environment may be missing from the underlying datasets.

Airspace restrictions are another important consideration. Although Skyplanr takes airspace information into account where available, its coverage varies by country. Users should always check the relevant national aviation authority’s maps and airspace information before flying. The generated flight plan should therefore be treated as a planning aid, not as confirmation that a flight is permitted or safe to perform.

Free to try

Skyplanr is available through a web browser and can be tried for free. Users receive three AI-generated flight plans per month without having to provide credit card details.

The paid Pilot subscription costs €24 per month, excluding VAT, and includes ten AI-generated plans. Additional plans can be purchased individually for €3 each.

The software supports DJI Fly, DJI Pilot 2 and Litchi, as well as exports to formats including KMZ, KML and GeoJSON.

Wiebe de Jager

Wiebe de Jager is the founder of Dronewatch (available in Dutch and English). Wiebe is an experienced drone pilot (EASA Specific category certified) and has published a number of bestselling books about drone photography and cinematography.

Leave a Reply

Sign up for our Newsletter