Tool

A reusable analysis operation packaged as an item, runnable on demand against caller-supplied inputs.

advanced

A tool is an item that wraps a single named analysis operation with a documented signature: declared inputs, declared parameters, declared outputs. Run it from its detail page, from a web app, or from another tool's pipeline.

Tools are the right item type when the same operation is run many times against different inputs and you want it to live in the catalog as a citable, shareable thing.

Tool vs. derived layer

  • Derived layer. Produces a specific output layer from specific inputs. The layer IS the result.
  • Tool. A reusable operation. Runs against caller-supplied inputs. Each run produces a new output (or modifies an existing one).

If you find yourself cloning the same derived layer with different inputs, the operation should probably be a tool.

What's stored

  • A name and description.
  • Input slots, each with a declared type (data layer, derived layer, geo boundary, single feature, etc.) and whether required.
  • Parameter slots, each with a type (number, string, choice list, distance, etc.) and a default.
  • Output declaration. What the tool produces (a new derived layer, a row count, a file, a side effect on an existing item).
  • The implementation. Either a built-in operation reference (one of the analysis steps) OR a script wrapped in the tool runtime.

Built-in tools vs. custom tools

The portal ships with built-in tools for the common cases (each analysis step is exposed as a tool). Custom tools let you bundle a multi-step pipeline with named parameters into one runnable unit.

A custom tool's implementation today is a derived-layer pipeline template: the same step types, just parameterized by tool inputs. A scripting hook (Python or JS) is planned but not in v1.

Running a tool

The detail page has a Run form. Fill in the inputs and parameters, click Run. The output appears (new item created, or existing item updated). Runs are logged to the tool's history tab.

Sharing

Standard three-tier. A public tool is runnable by anyone with the link; an org-only tool is gated.

Notes

  • Tools are not webhooks. They run synchronously in a request. If you need scheduled or event-driven runs, schedule them via the admin scheduler.
  • Quotas. Org admins can set a per-tool runtime limit and a per-user concurrent-run cap.