field.tours
IVAR Studios
In developmentA 3D scan records a place. field.tours lets people visit it - and talk to its recorded history. field.tours streams 3D scans of any size into a browser, a touch display or a VR headset, tags what is in them and connects them to your body of scientific data. An AI host takes visitors through the place and answers every question from your own sources. Scans that sat unused on a hard drive become places people can explore, question and learn from.
Visit field.tours →
A scan shows what a place looks like. field.tours adds what each part of it is. These tags let the AI host point at a storehouse, fly to a rune, take you around a ship and tell you where each fact comes from.
One engine, any place
From a single runestone to a glacier of about 25 square kilometres: every field.tours story runs on the same AI-powered engine. The engine controls the 3D world, the camera, the stops, the AI host, the voice, the two languages and VR. A story only supplies the place and the text about it.
field.tours runs in a regular web browser. It also runs full screen on a touch display in a visitor centre, and in a VR headset. Visitors do not download anything. They select who the host speaks to, from school pupils to specialists, and the host adapts its replies.
Semantic tagging of the 3D world
To a computer, a scan is only shapes. So we tag the important parts of the 3D world. On a runestone, we can tag each individual rune or relevant marking. On a glacier landscape, we can tag what is the ablation zone and what is the moraine - so that the visitor can ask about anything, and the AI host knows what to show them.
The AI host uses these tags to understand the world. With them, it can point at a part, light it up, fly you to it and see what is on your screen. Ask "what am I looking at?" and it knows. Ask where a fact comes from, and you open the source with one tap. If the material does not say something, the host tells you so. It does not guess.
Custom stories with field guides
Sometimes a story needs more than the place. Then we write and build it with you, on the same engine. A custom story can add animation and simulation, such as a drone that surveys, a boat that lands or ice that falls, and interaction made for what your story asks of the visitor.
It can also give the AI host a body. We call it a field guide: a character in the scene, such as a drone over a glacier or an underwater vehicle in the sea. It flies between the stops, turns to face you and flies out to point at what it talks about. Ask it to come closer or fly a survey pass, and it does. It speaks each reply and shows it as subtitles, and you can type or speak to it. At Sálajiegna, a glacier front that no visitor can approach, SAL-01 takes you through the ice.
A field guide never pretends to be a person. It is a guide program based on the research, and the findings belong to the scientists. Ask it something the study does not cover, and it tells you. It does not guess. We can design a new field guide for your story, with its own body, its own voice and knowledge from your sources.
The host knows what is on screen
Go anywhere in a tour and ask "what am I looking at?". The host reads the view and finds the tagged parts that are on your screen. Then it answers about the building, object or feature that is in front of you. You can also tap anything that has a name, and the question is written for you.
The stops make a route, but you can leave it. You move freely, and the host follows. When none of its tagged parts is in view, it tells you so.
Live connections to your own records
The host can also use live data, not only fixed texts. We can connect a tour to your own databases, for example your collection on DigitaltMuseum or an internal catalogue. Ask about an object, and the answer comes from the live record. When you correct the record, the tour shows the correction on the next visit.
The same type of connection can go to a sensor feed or an operator's live data. This makes the engine as useful for a power station as for a museum.
Sources, not guesswork
The host only uses the tour's own sources: your texts, your collection records, research papers and the reference works you choose. Ask where a fact comes from, and the host tells you. Each stop also lists its sources, and you can open and read them.
The host never fills gaps from its general training. When the sources disagree, the tour says so. When they say nothing, the host says that it does not know. This is why an organisation can safely put its name on a tour.
Streamed 3D
A scan of a real place is large. One of our sites is 2.1 GB as delivered and has 194 million triangles. To open it, field.tours downloads 12 MB. The engine streams the scan as 3D tiles. It sends only the part you look at, at the detail your screen can show. The rest arrives as you move.
Visitors do not see this process. They see the finished scene from the first second: on the web, on a touch display and in VR. They install nothing, and nothing is left out. Give us a raw scan, a point cloud, a photogrammetry model or a BIM file. Our pipeline turns it into a world that streams.
Become our pilot partner
A story in field.tours is content, not code, so your own staff can change a stop, add a language or correct a source. We build a Standard tour from your model and texts, or write a Custom story with you. We are looking for a few museums, companies and research groups to test it on a real place.
Become a pilot partner →
Impact
We believe the impact on learning can be immense. When a real place comes to life in interactive 3D, and every answer rests on accurate scientific data, people can explore places they will never visit and ask the questions they actually have. A school class can stand at a calving glacier front. A visitor can read a runestone rune by rune.
We build this with care, especially for young learners. Read our stance on AI for learners.
Credits
- Swedish National Maritime and Transport Museums - 3D model of Baron Anders von Höpken (1759), CC BY 4.0
- Matt Maksymowicz - "DJI Mavic 2 Pro" 3D model, the body of SAL-01, CC BY 4.0
- Vacek et al. (2024) - The 2022 Sálajiegna field campaign behind the glacier scan and SAL-01's knowledge
- Google and Cesium ion - Photorealistic 3D Tiles and terrain around the scans