Create Accountor Login
Docs

Datasets

Part of the MountDev Ops Workspace documentation. Updated November 16, 2026

MountDev Ops Workspace documentation | Updated November 16, 2026

A Dataset is a named list of rows with typed fields, for outside data you want in your workspace: orders from your shop, readings from a device, numbers from another system. You make a Dataset, send it rows, and use it as a source for a custom widget.

Our Commitment

MountDev Ops is built security first. Your business data belongs to you, and you can export all of it at any time. We never use analytics or tracking of any kind inside MountDev Ops.

Every update passes all 5,581 automated checks before it ships, and every screen change is walked through before release.

Making a Dataset

Owners and admins open Settings, Your Own Modules, then Datasets, and choose New Dataset (action core.datasets.create).

  • Name and a short key, made from the name unless you give one (lower case letters, digits, and underscores). The key is how a Workflow, the API, and the MCP name the Dataset.
  • Fields, each with a name and a type: Text, Number, Money (in dollars, kept as whole cents), Date (such as 2026-03-14), or Yes or No. Up to 30 fields.
  • Private. Each field is Private when you make it, and you choose otherwise one field at a time. A Private field is never shown to your clients or to the public, whatever a widget does with the Dataset.
  • Who May Use It. Owners and admins always may. You can also allow every member, and any teams (Groups) you choose. Someone who may not use a Dataset does not see it listed and cannot read it or add to it.

Change (action core.datasets.edit) renames a Dataset, adds or removes fields, changes whether a field is Private, and changes who may use it. A field keeps its type once the Dataset has rows; add a new field instead. A field you remove from the list keeps its values in the rows.

Rows

Rows shows a Dataset's newest rows (action core.datasets.read, up to 100 at a time). Add Row adds one by hand, and Remove (owners and admins, action core.datasets.rows.remove) takes one away.

A row can carry an outside id: the id the row has in the other system. Sending a row with an id that is already there updates that row, so it is never doubled. A row with no outside id is always new. A Dataset holds up to 100,000 rows.

Rows are quiet, like an import: adding one never starts an automation, a Workflow, a webhook, or a notice. Every row is in your own workspace, in your own region, and in Export Everything (Datasets.csv, and one file for each Dataset with its rows, money in dollars). Nothing about a Dataset is visible to anyone outside your company.

Bringing Outside Data In

Outside data comes in by the tools every account has. There are three ways, and they all use the one action, core.datasets.rows.add.

From a Workflow That Receives a Webhook

  1. In Workflows, make a workflow that starts from another service (the signed address a workflow gets), and pick the fields you want from what it sends, such as order.number and order.amount.
  2. Add the step Add a row to a Dataset (core.dataset_add_row). Give it the Dataset's key, an outside id (for example the order number), and the values, using the picked fields: outside_id is {{input.order}}, and values is { "order_number": "{{input.order}}", "amount": "{{input.amount}}" }.
  3. Turn the workflow on. Each delivery adds a row, and a repeat of the same order updates it.

From a Workflow That Fetches on a Schedule

  1. Connect the other service in Settings, Connections, Connected Services.
  2. Make a workflow that runs on a schedule, with the step Call a web service first, then Add a row to a Dataset using what the call handed on: outside_id is {{steps.call.site_id}} and values uses {{steps.call.count}}.
  3. Turn it on. Each run fetches and adds or updates the rows.

From the API or the MCP

Send a row with your account's API (a key made in Settings, Connections, API Keys) or through an AI agent connected by the MCP. The action is core.datasets.rows.add with the Dataset's key, an optional outside id, and the values by field key. An example body: { "dataset": "shop_orders", "outside_id": "A-1001", "values": { "order_number": "A-1001", "amount": "1250.50", "ordered": "2026-03-14", "shipped": true } }. A key or agent acts with the access of the person behind it, so it can add to a Dataset only when that person may use it. Making and changing a Dataset, and removing rows, are for people only.

A workflow step runs as the person who turned the workflow on, with their access at the time, and its Activity Log line names the workflow and the run. The run keeps ids and counts, never the values.

Reading the Datasets

An AI agent or API key can list the Datasets it may use (core.datasets.list) and read their rows (core.datasets.read), the same as the person behind it. Fixing or cleaning outside data is the workflow's job: a Dataset keeps what it is sent.

Ask Us Anything

We’d love to hear from you!