Demand to add the pgPointcloud extension to Supabase

[sblaisfernandez]

Hello Supabase team!

I'm a user of your product, and I'm exploring geospatial application. I would like to add point cloud to my Postgres/PostGIS database on supabase.

It would be really nice if you guys can add the extension pgpointcloud allow the manipulation of LiDAR data.

Thanks, I really enjoyed the product.

[sblaisfernandez]

In the same direction, It would be nice to add the extensions timescaledb. Thanks

[harrysun2006]

Glad you're enjoying it! pgPointcloud would be a nice addition for LiDAR work, but it isn't in the managed extension list today, and you can't add arbitrary extensions yourself on hosted Supabase (only the ones Supabase has vetted and packaged), so this is genuinely a feature request for the team rather than something you can enable.

Worth keeping the request open and getting other geospatial folks to upvote — pgPointcloud pairs naturally with PostGIS (which is available), so there's a coherent story for it.

In the meantime, depending on what you're doing with the point clouds, a couple of interim options:

• If you mainly need to store and query points spatially, PostGIS geometry(PointZ) / MultiPointZ covers a lot of ground (indexing, spatial filters, nearest-neighbour) even if it's not as compact as PcPatch.
• If you need the heavier LiDAR processing (patch compression, PDAL pipelines), do that outside the DB and land the results — derived features, tiles, or PostGIS geometries — into Supabase.

Not as clean as native pgPointcloud, but it keeps you moving while the extension request is under consideration.

[yakubka]

pgPointcloud is not currently available on Supabase Cloud. PostGIS is supported but pgPointcloud requires additional shared libraries compiled into Postgres that the managed platform does not include at the moment.

For LiDAR work on Supabase today there are two practical paths.

The first is to store the raw point cloud data in Supabase Storage as LAZ or LAS files and process them externally with a tool like PDAL or libLAS. You keep the metadata and derived attributes (bounding box, point count, classification statistics) in a Postgres table alongside a PostGIS geometry column for spatial indexing, and retrieve the actual point data from Storage on demand. This pattern works well for read-heavy workflows where you query on metadata and then fetch the raw file.

The second is to self-host Supabase with a custom Postgres image that includes pgPointcloud. The self-hosting documentation at supabase.com/docs/guides/self-hosting describes how to extend the Postgres image. You would build pgPointcloud from source, add it to the image, and run it alongside the standard Supabase stack. More operational overhead, but you get the full extension.

I would encourage filing a GitHub issue requesting pgPointcloud as a managed extension. The more concrete use cases and upvotes an extension request has, the easier it is for the team to prioritize adding it to the managed platform.