Download a piece of the world
for you (or your LLM)

Aleph does two things:

• Lets you draw a rectangle or polygon on the map and downloads everything inside it (buildings, elevation, HD satellite imagery, Street View photos, textured 3D models).
• Provides any of the above information for specific points on Earth, on the fly, via the CLI.

This was born as a way to let LLMs quickly gather information from the physical world, but you can just use it to have piece of the Earth on your laptop.

Caution

Data licenses:

• Street View, satellite imagery, 3D meshes: © Google and its providers. Downloading them goes against Google's terms. Not affiliated with Google. You are responsible for how you use this tool and the data it downloads.
• OSM data and geocoding: ODbL. Commercial use allowed with "© OpenStreetMap contributors" credit. Shared modified databases must stay ODbL.
• Terrain: AWS Terrain Tiles, open sources. Commercial use allowed with the credits listed there.

Also, this tool uses undocumented APIs so it may break or hit rate limits.

Install

Requires Python 3.11+. The only dependencies are Pillow, numpy, and truststore. No API keys are required.

git clone https://github.com/Belluxx/Aleph.git
cd Aleph
python3 -m venv .venv
source .venv/bin/activate
python -m pip install .

Then test with

alephgeo --help

Open the dashboard

alephgeo dashboard -o captures

The dashboard opens in your browser at http://127.0.0.1:8100. Draw a rectangle or polygon on the map, choose the sources and their settings, check the preview and time estimate, then start the capture. Open any capture to show or hide each layer, or to explore its 3D mesh or terrain relief in 3D. Captures are saved in captures.

The basemap comes from OpenFreeMap and needs an internet connection.

Use the CLI

Download a Street View photo or satellite image by place name:

alephgeo streetview --place "Colosseum, Rome" --best-match -o captures
alephgeo satellite --place "Colosseum, Rome" --best-match -o captures

--best-match selects the first match. Each download is saved in a new folder inside captures.

See the CLI docs for examples and details.