Getting started

A delivery to GEORIZON has three stages.

1. A dataset is created in the GEORIZON web app, with a dataset type that decides how it will be read, and a coordinate reference system. This determines the contract your data has to meet.
2. You upload into the dataset's bucket, using the folder layout that dataset type expects. Access is over S3-compatible credentials — see the dataset's Files view in the web app, or make an access key under Profile settings › S3 keys.
3. Processing runs. What the delivery holds is indexed; derived artefacts (COPC point clouds, deep-zoom tiles, blurred imagery) are generated from there. Failures surface on the dataset's workflow page with the step that failed and why.

The rest of these docs are about stage 2, because that's the stage where the requirements are yours to meet.

An assistant connected to the MCP server can create the dataset, submit the delivery and read the validation report. Uploads still go through S3 with an access key from your profile, not through MCP.