Web UI

Build, run and share pipelines in the browser. No code, no install.

The web UI is Spade for people who would rather not write YAML. Everything happens in the browser: find a block, drop it on a canvas, wire it up, run it in the cloud, and look at what came out. Pipelines built here are the same pipelines the CLI runs, so a team can mix both.

Build pipelines on a canvas

The pipeline editor showing a terrain pipeline: two data.read blocks feed gdal.clip_raster_by_vector, which feeds gdal.hillshade and gdal.slope, which feed viz.raster_map and gdal.raster_histogram.
A terrain pipeline in the editor, laid out by dependency.
  • Wiring that works itself out. Connect two blocks and the editor matches outputs to inputs by type. When more than one match fits, it asks, and shows each input's and output's description so the choice is an informed one.
  • Settings as forms. Each block's arguments appear as a form built from its manifest, with defaults, allowed values and descriptions, plus a picker for any secrets it needs.
  • Problems before you run. Missing arguments and unresolved connections are listed as you build. When you press Run, the pipeline is checked against the real block definitions before anything is queued.
  • Undo, arrange, import and export. Lay a pipeline out by dependency in one click, undo mistakes, and move between the canvas and YAML in either direction.

Find the right block

The block browser searching for hillshade, with filters for data type, task, collection, accepted and produced formats.
Search by what a block does or the data it takes. Every filter shows how many blocks it would leave.

The catalogue is searchable by name, description, tags, and the formats a block accepts and produces. Searching for crop also finds clip, and searching for LiDAR also finds point cloud. Filters narrow by data type, task, collection, format and requirements, and the whole view lives in the URL, so a filtered list can be shared as a link.

The gdal.hillshade block page listing its settings, inputs, outputs, requirements and an example pipeline fragment.
Each block has a page with its settings, inputs and outputs, requirements, and a ready-to-paste example.

Run it and see what came out

A finished run showing a slope histogram as a table, a slope map image, and the logs of each block.
A finished run: outputs previewed in place, and each block's logs under its own name.

Runs execute in the cloud, in parallel where the pipeline allows. When one finishes:

  • Outputs preview in place. Images, tables, charts and reports open on the run's page. Every file downloads, singly or as a ZIP.
  • Logs by block. Each block's output is listed under the block that wrote it, so a failure points straight at its cause.
  • Storage you can see. The page shows what a run's outputs take up, and the owner can delete them once they are no longer needed.
The run history listing three runs, two succeeded and one failed, with the storage they use.
Run history, newest first, with each run's status.

Bring your own data

The data page listing uploaded files: plots.csv, county_estimates.csv, dem.tif and study_area.geojson.
Uploaded files, ready to use in any pipeline.

Upload files or whole folders, including large geospatial ones, and use them in pipelines like any other input. An interrupted folder upload can be resumed. Data is shared without uploading it again.

Share and reuse

The pipelines page showing three saved pipelines with Open, Delete and Share buttons.
Saved pipelines, each one a click from the editor.

Pipelines, results and data can each be kept private, shared with named colleagues by email, or made public. A public pipeline appears in the library for anyone to open and run. Collections of blocks can be private too, and shared with the people who need them.

Prefer to just ask?

An AI assistant connected to Spade can find blocks, draft a pipeline, check it and run it for you, and every run it starts shows up here.

Prefer the terminal?

The command-line interface is the developer side of Spade: write blocks, run pipelines locally, and publish collections. The two share one pipeline format: a pipeline exported from the editor runs with spade run, and one written by hand imports onto the canvas.

← All Features