Embeddable photogrammetry viewer for the web
With our LOD-streaming approach, Sangine lets you embed high-fidelity photogrammetry scenes with 100+ M triangles and 70+ 4K textures directly on your pages.
You can also share your scenes privately for inspection and collaboration.
The Burden of Dreams, the worlds hardest boulder problem rendered with 30+ M triangles
Features
Embeddable
Drop a single line of code into any webpage and your model is live — fully interactive, no plugins, no friction.
Quality
Our streaming renderer delivers your photogrammetry captures at their full captured density — directly in the browser, exactly as scanned.
Private sharing
Share models with clients or collaborators via a private link — no account required on their end, full control on yours.
Our story
Santeri Oksanen, founder
Sangine was born as a hobby project. I got into photogrammetry based 3d-scanning while studying robotics and computer vision at Aalto University, and wanted to try it with my long-standing hobby - outdoor climbing. After scanning a few boulders nearby, I quickly noticed that to inspect the holds - which may be 5 mm wide - in a useful way in the scanned model, the rendering requires extremely dense meshes and high texture counts. Practically triangle count easily reaches tens of millions of triangles for a single boulder, while the required texture-count for visual fidelity easily reaches 100 4k textures. No existing web-based renderer seemed to solve this out of the box.
The extreme triangle and texture count required splitting the mesh into spatially coherent pieces, and also resampling the textures. After building this preprocessing pipeline, I was genuinely impressed by the rendered quality, I could see real details even on the smaller holds. I released the first publicly available site in spring 2025 on crags3d.sanox.fi, with some really promising feedback, although the initial render was still fairly slow thanks to suboptimal mesh divisions, and a single texture spanning multiple mesh parts - making LOD streaming practically really hard. At this point, my ex-colleague joined the project.
After the initial positive feedback, we fixed the partitioning issues, fixed the flat background with a clear sky, and added lighting calculations (reflections, shadow maps, ambient occlusion) to the rendering engine. At this point we felt it was time to give it a go. So we put the renderer and the preprocessing pipelines together to form Sangine. Now we are curious to see where this technology can go beyond climbing.
Pricing
- Maximum one model
- Branding in embeds
- Only public scenes
- 100 GB storage space
- Unbranded embeds
- Private scenes and shares
How it works
-
1
Upload your archive
Package your scan as a ZIP containing an
.obj,.mtl, and textures. That's your only input. -
2
Automatic preprocessing
The system partitions the mesh into spatially coherent clusters and resamples textures for streaming. Runs in the background — no manual steps required.
-
3
Configure the view
Open the scene editor to set the initial camera position and fine-tune lighting — sun altitude, azimuth, color temperature, and intensity.
-
4
Share or embed
Generate a private share link or copy a single line of HTML to embed the interactive viewer anywhere. No account needed for viewers.
Asset preprocessing
The preprocessing pipeline partitions the mesh into spatially coherent clusters and resamples textures per cluster. This spatial decomposition is what enables efficient LOD streaming — only the geometry visible from the current camera position is loaded and rendered.
Adaptive quality rendering
The renderer evaluates the viewer's hardware capabilities and selects appropriate quality settings at startup. As the camera moves, it continuously adjusts which LOD levels are streamed per cluster — keeping frame rates smooth on modest hardware while delivering full fidelity on capable devices.
Sign up
Get early access
Leave your details and we'll be in touch.
Early access
Sangine is currently in early access — a limited number of spots are available. Tell us about your use case and we'll be in touch to set up your account.
-
1
Fill in the form
Tell us your email, use case, and how much storage you expect to need.
-
2
We get in contact
We'll reach out to set up your account and answer any questions.
-
3
Upload and embed
Once your account is ready, upload your models and embed them anywhere with a single line of code.