Build your first workflow

In the Anythink dashboard

1. Go to Settings › Data Model and select Add Entity.
2. Enter orders as the Entity Name and select Create Entity.
3. Open the entity and select Add Field for customer_email (Small text), total (Decimal) and status (Small text).
4. Repeat for a second entity, order_reviews, with order_id (Integer) and note (Large text).

With the CLI

anythink entities create orders
anythink fields add orders customer_email --type varchar
anythink fields add orders total --type decimal
anythink fields add orders status --type varchar

anythink entities create order_reviews
anythink fields add order_reviews order_id --type integer
anythink fields add order_reviews note --type text

With an AI assistant (MCP)

Create an orders entity with customer_email (text), total (decimal) and status (text). Then create an order_reviews entity with order_id (integer) and note (text).

The assistant runs entities create and one fields add per field through the cli tool. See Model your data for every field type.

In the Anythink dashboard

1. Open Workflows and select New Workflow.
2. Enter the Name flag-large-orders.
3. Under Triggers, leave Trigger Type as Event. Set Event Type to Entity Created and Entity to orders.
4. Under Fire only when…, select Add filter. Set the field to total, the operator to ≥ and the value to 100.
5. Leave Enable Workflow cleared for now, then select Create Workflow.

The editor opens with a Start node on the canvas.

With the CLI

anythink workflows create flag-large-orders \
  --trigger Event \
  --event EntityCreated \
  --entity orders \
  --filter '{"field":"total","op":"gte","value":100}' \
  --enabled

✓ Workflow flag-large-orders created (id: 42).

Note the id it prints. The CLI enables the workflow as it creates it. Nothing runs until an order arrives, and you add the step next.

With an AI assistant (MCP)

Create a workflow called flag-large-orders that runs when an order is created and its total is at least 100, and enable it.

The assistant runs workflows create with an event trigger and a filter through the cli tool, and tells you the workflow's id.

In the Anythink dashboard

1. In the editor, select Add Step and choose Create Data.
2. Enter the Name Create review and the Key create_review, and tick Enable Step.
3. Set Entity to order_reviews.
4. Set the Payload:

{
  "order_id": "{{ $anythink.trigger.id }}",
  "note": "Order from {{ $anythink.trigger.data.customer_email }}, total {{ $anythink.trigger.data.total }}, needs review"
}

5. Save the step. On the canvas, drag from the Start node to the step's input to make it the start step.
6. Select Save Workflow, then select Enable Workflow in the toolbar.

With the CLI

In a CLI step the payload is a string that holds JSON, so its inner quotes are escaped. Use the workflow id from the previous step:

anythink workflows step-add 42 create_review \
  --name "Create review" \
  --action CreateData \
  --start --enabled \
  --params '{"entity_name":"order_reviews","payload":"{\"order_id\":\"{{ $anythink.trigger.id }}\",\"note\":\"Order from {{ $anythink.trigger.data.customer_email }}, total {{ $anythink.trigger.data.total }}, needs review\"}"}'

✓ Step create_review (id: 82) added to workflow 42.

--start makes it the step the job begins with.

With an AI assistant (MCP)

In workflow 42, add a Create Data step called create_review that writes to order_reviews, and make it the start step.

The assistant runs workflows step-add through the cli tool. Then open the step in the Anythink dashboard and add the two templates from the dashboard steps above to its payload.

In the Anythink dashboard

1. Open the orders entity and select Create Order.
2. Enter sam@example.com, a total of 149.50 and status new, then save.
3. Create a second order with ali@example.com, a total of 24 and status new.

With the CLI

anythink data create orders --data '{"customer_email":"sam@example.com","total":149.50,"status":"new"}'
anythink data create orders --data '{"customer_email":"ali@example.com","total":24.00,"status":"new"}'

With an AI assistant (MCP)

Create two orders: one for sam@example.com with a total of 149.50 and status new, and one for ali@example.com with a total of 24 and status new.

The assistant runs data create twice through the cli tool.

In the Anythink dashboard

1. Open the workflow and select View Job History.
2. There is one job, for the 149.50 order, with the status Success. Open it to see the Payload it received and each step under Execution Steps.
3. Open the order_reviews entity. There is one record: "Order from sam@example.com, total 149.5, needs review".

With the CLI

anythink workflows jobs 42
anythink data list order_reviews --json

The first command lists the workflow's jobs and their statuses: one, with status Success. The second shows the review the step created:

[
  {
    "id": 1,
    "order_id": 1,
    "note": "Order from sam@example.com, total 149.5, needs review"
  }
]

The output also includes created_at, updated_at and a few other system fields.

With an AI assistant (MCP)

List the order_reviews records.

The assistant runs data list order_reviews through the cli tool. Open the workflow's View Job History in the dashboard to see each job.

The 24 order didn't match the filter, so it left no job at all. To see a job's step log in full, use the dashboard's View Job History. Test workflows and read job history covers reading jobs, running a workflow by hand and tracking down failures.