<img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/assets/logo/forge3d_dark.svg" alt="forge3d" width="320"> GPU-accelerated 3D terrain rendering for Python. Built in Rust with WebGPU. Pre-built wheels — no Rust toolchain required. <a href="https://pypi.org/project/forge3d/"><img src="https://img.shields.io/pypi/v/forge3d?color=EFA026&style=flat-square" alt="PyPI"></a> <a href="https://pypi.org/project/forge3d/"><img src="https://img.shields.io/pypi/pyversions/forge3d?color=D0C8BA&style=flat-square" alt="Python 3.10+"></a> <a href="https://github.com/milos-agathon/forge3d/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-Apache--2.0%20%2F%20MIT-blue?style=flat-square" alt="License"></a> <a href="https://milos-agathon.github.io/forge3d/"><img src="https://img.shields.io/badge/docs-online-blue?style=flat-square" alt="Docs"></a>

This image was generated with 5 lines of Python. Load a DEM, launch the GPU viewer, position the camera, and capture a publication-quality snapshot. No shaders to write. No OpenGL boilerplate. Just pip install forge3d.

import forge3d

with forge3d.open_viewer_async(terrain_path="rainier.tif") as viewer:
    viewer.set_orbit_camera(phi_deg=225, theta_deg=35, radius=1.2)
    viewer.set_sun(azimuth_deg=315, elevation_deg=32)
    viewer.snapshot("render.png", width=3840, height=2160)

Install

pip install forge3d

Need Jupyter widgets or tutorial datasets?

pip install "forge3d[jupyter]"     # ViewerWidget for notebooks
pip install "forge3d[datasets]"    # on-demand sample DEMs, CityJSON, COPC
pip install "forge3d[all]"         # everything

Gallery

<table> <tr> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/01-mount-rainier.png" width="280"> PBR terrain</td> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/03-swiss-landcover.png" width="280"> Landcover overlay</td> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/06-point-cloud.png" width="280"> LiDAR point cloud</td> </tr> <tr> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/09-shadow-comparison.png" width="280"> Sun & shadow control</td> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/02-mount-fuji-labels.png" width="280"> GeoPackage labels</td> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/05-3d-buildings.png" width="280"> CityJSON buildings</td> </tr> <tr> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/04-luxembourg-rail-network.png" width="280"> Vector overlays</td> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/07-camera-flyover.png" width="280"> Camera animation</td> <td align="center"><img src="https://raw.githubusercontent.com/milos-agathon/forge3d/main/docs/gallery/images/10-map-plate.png" width="280"> Map plate compositor</td> </tr> </table> <a href="https://milos-agathon.github.io/forge3d/gallery/">See the full gallery →</a>

What You Get

Open Source (Apache-2.0 / MIT)

Everything you need to go from raw elevation data to a rendered 3D scene:

Interactive viewer — real-time orbit, pan, zoom via a Rust/WebGPU subprocess controlled from Python over IPC
Terrain rendering — load GeoTIFFs or numpy arrays, PBR materials, 100+ colormaps
Vector overlays — GeoJSON/GeoPackage polygons, lines, and labels projected onto terrain
Point clouds — COPC and LAZ files with millions of points, colored by elevation or classification
3D Tiles — stream OGC 3D Tiles tilesets directly into the viewer
CRS reprojection — automatic coordinate transforms via PROJ + pyproj
Camera animation — keyframed flyover paths with frame-by-frame export
Jupyter integration — ViewerWidget embeds the viewer inline in notebooks
High-res snapshots — up to 8K PNG export from any camera angle

Pro

Professional cartography and production workflows:

Map plate compositor — legends, scale bars, north arrows, multi-panel layouts
SVG / PDF export — publication-ready vector output
3D buildings — GeoJSON, CityJSON, and 3D Tiles import with roof inference and PBR materials
Mapbox Style Spec — load and apply Mapbox-compatible styles
Scene bundles — save and share complete .forge3d scene packages
Commercial licensing details → https://github.com/milos-agathon/forge3d#license

Tutorials

Two tracks depending on your background:

GIS professionals — Visualize your first DEM →
Python developers — Your first 3D terrain →
Architecture overview — How forge3d works →

Jupyter

from forge3d.widgets import ViewerWidget

widget = ViewerWidget(terrain_path="dem.npy")
widget.set_camera(phi_deg=225, theta_deg=35, radius=1.2)
widget.set_sun(azimuth_deg=315, elevation_deg=32)
widget.snapshot()  # renders inline

Links

Documentation · API Reference · GitHub · PyPI

License

Open-source core released under Apache-2.0 OR MIT. Pro features require a commercial license key set with forge3d.set_license_key(...). Licensing details live at https://github.com/milos-agathon/forge3d#license

Forge3d

Install / Use

README