Internal tools and dashboards

In the Anythink dashboard

1. Go to Settings › Data Model and select Add Entity. Create employees, tick Add to Menu if you want it in your sidebar, and add its fields.
2. Create expenses and add its fields from the table. For each select field, choose the Select display and add the options. For employee, choose Many-to-One and pick employees.
3. Open the expenses records and select Import, then open New Import.
4. Drop your .csv file on the upload area. The limit is 50 MB.
5. Select Auto-Match, fix any headings it missed, and select Run Import. Here Description maps to title, Cost to amount, Type to category, State to status and Date to spent_on.
6. Follow progress in History. Select View Logs for the log.

With the CLI

anythink entities create employees
anythink fields add employees name --type varchar --required --name-field
anythink fields add employees department --type varchar --display select --options "sales,ops,engineering"

anythink entities create expenses
anythink fields add expenses title --type varchar --required --name-field
anythink fields add expenses amount --type decimal
anythink fields add expenses category --type varchar --display select --options "travel,software,meals"
anythink fields add expenses status --type varchar --display select --options "submitted,approved,rejected" --default submitted
anythink fields add expenses spent_on --type date
anythink fields add expenses employee --type many-to-one --target employees

The CLI has no import command. Upload the file, then start a job with a mapping from each heading to a field:

anythink files upload expenses.csv

anythink fetch /imports/jobs --method POST --body '{
  "entity_name": "expenses",
  "file_id": 4226,
  "mapping_json": "{\"Description\":\"title\",\"Cost\":\"amount\",\"Type\":\"category\",\"State\":\"status\",\"Date\":\"spent_on\"}"
}'

anythink fetch /imports/jobs/2

file_id is the id files upload prints. The last call returns the job's status, processed_records and failed_records.

With an AI assistant (MCP)

Create employees with a name and a department select field, and expenses with a title, a decimal amount, select fields for category and status, a spent_on date, and a many-to-one employee field. Then upload expenses.csv and import it into expenses, mapping Description to title, Cost to amount, Type to category, State to status and Date to spent_on. Tell me how many rows failed.

The assistant runs entities create and fields add, then files upload and the import endpoints, through the cli tool. Connect Claude to your backend with MCP sets up the MCP server.

An import only creates records. It never updates existing ones, so importing the same file twice creates every row twice. A row the database rejects is skipped and the rest still load. Records, views and CSV import covers the file rules, saved mappings and what a failed row does. For every field type, see Model your data.

In the Anythink dashboard

1. Open the expenses entity from the sidebar. Select a column heading to sort. Open a column's menu to filter it.
2. Select Kanban View, then choose status in the Status Field menu. Drag a card to another column and the record updates at once.
3. Select Calendar View, and switch between Month, Week, Day and Agenda. Select an event to open its record.

With the CLI

Views live in the dashboard. The CLI works on the records behind them:

anythink data list expenses --filter 'status=submitted' --limit 10
anythink data update expenses 4 --data '{"status":"approved"}'

The second command is what dragging a card to approved does.

With an AI assistant (MCP)

Show me the submitted expenses over £100, then mark expense 4 as approved.

The assistant runs data list and data update through the cli tool. Open the entity in the Anythink dashboard and switch to Kanban View or Calendar View to see the result.

The kanban board and calendar load up to 1,000 records. The calendar picks its date field for you from the field names, so call yours start_date and end_date if you need control. The grid remembers your layout, but only in that browser. See Records, views and CSV import for the details.

In the Anythink dashboard

1. Open Settings › Users and select the Roles & Permissions tab. Select +, enter Finance, tick API Access and select Create Role.
2. Select the role, tick the permissions from the table and select Update Role.
3. Open Settings, then Menu Configuration, and select Add Menu. Name it Finance and choose the Finance role.
4. Select the menu, then Add Menu Item. Choose Entity, pick expenses and choose an icon. Do the same for employees.
5. To add a dashboard to the sidebar, add a Custom Link with the path /dashboard/<id>, using the dashboard's id from your browser's address bar.
6. Invite people under Settings › Users with the Finance role.

With the CLI

anythink roles create Finance --description "The finance team"
anythink roles list
anythink menus list
anythink menus add-item 12 expenses --icon Receipt
anythink menus add-item 12 employees --icon Users

Replace 12 with the id of the Finance menu, from anythink menus list.

anythink users invite maya@example.com Maya Patel --role-id 104

With an AI assistant (MCP)

Create a role called Finance, then add expenses and employees to the Finance menu with the Receipt and Users icons. Invite maya@example.com, Maya Patel, as a Finance user.

The assistant runs roles create, menus add-item and the invitation through the cli tool. Role permissions and creating the menu are dashboard tasks, and the assistant can walk you through them.

When Maya signs in, her sidebar shows Dashboard, Expenses, Employees and Help. Administrators keep the full sidebar.

A menu item is a shortcut, not a permission. Hiding a link doesn't block access, and showing one doesn't grant it, so set up both. See Menu configuration and Roles and permissions. If one team should only see its own records, add row-level security: see Groups and row-level security.

In the Anythink dashboard

1. Open Dashboard in the sidebar. Open the dashboard switcher and select New dashboard. Name it Finance, turn on the share switch so the team can see it, and create it.
2. Select the pencil, then Add widget and choose Chart. Select Design a new chart.
3. Pick expenses as the data source and Pie as the chart type. Set the group-by field to category, the aggregate to count, and the timeframe to Last 30 days. Save it.
4. Add another Chart widget. Choose Bar, set the X axis to spent_on with a week bucket, the aggregate to sum and the Y field to amount.
5. Select Add widget, choose Materialised view, and name it Employee expenses. Under Show one row per, pick employees.
6. Select Add column. Set the Source to entity, the Entity to expenses and Match on to employee. Set the Display to exists and add a column filter, status equals submitted.
7. Add a second column the same way, with the Display set to value, the Value field set to amount and Order set to latest. Check the Preview and save.
8. Select Save on the dashboard.

With the CLI

Charts are saved through the REST API. Chart request bodies use camelCase names, and responses come back in snake_case.

anythink fetch /charts --method POST --body '{
  "name": "Expenses by category",
  "chartType": "pie",
  "entityName": "expenses",
  "xField": "category",
  "yAgg": "count",
  "timeframePreset": "30d"
}'

anythink fetch /charts --method POST --body '{
  "name": "Spend per week",
  "chartType": "bar",
  "entityName": "expenses",
  "xField": "spent_on",
  "xInterval": "week",
  "yAgg": "sum",
  "yField": "amount",
  "timeframePreset": "90d"
}'

Each response includes the chart's id. Put the charts on a shared dashboard with chart widgets:

anythink fetch /dashboards --method POST --body '{
  "name": "Finance",
  "is_shared": true,
  "widgets": [
    {"type":"chart","title":"Expenses by category","config_json":"{\"chartId\":40}","sort_order":0},
    {"type":"chart","title":"Spend per week","config_json":"{\"chartId\":41}","sort_order":1}
  ]
}'

Materialised views are built in the Anythink dashboard, so add the Employee expenses table there, with the steps above.

With an AI assistant (MCP)

Create a pie chart of expenses grouped by category over the last 30 days, and a bar chart of the sum of amount per week over the last 90 days. Then create a shared dashboard called Finance with both charts on it.

The assistant creates the charts and the dashboard through fetch. It can't build a materialised view: add that in the Anythink dashboard. Arranging and resizing widgets is also easier there.

A chart only counts the records its viewer is allowed to read, so with row-level security on, two people can see different numbers on the same dashboard. Charts and views are calculated each time they're opened, so they're always current. For a trend over time use a chart, and for "which of these records have a match elsewhere" use a materialised view. See Dashboards and charts for every chart type, the timeframe and compare settings and sharing, and Materialised views for columns and the TOTAL row.

In the Anythink dashboard

1. Open Workflows and select New Workflow. Name it expense-submitted.
2. Set Event Type to Entity Created and Entity to expenses. Select Create Workflow.
3. Select Add Step and choose Send An Email. Name it Email finance with the Key send_notice.
4. Enter finance@example.com as the To Email Address and pick the expense_submitted template.
5. On the Mappings tab, set title to {{ $anythink.trigger.data.title }} and amount to {{ $anythink.trigger.data.amount }}.
6. Save the step, make it the start step, save the workflow and select Enable Workflow.

With the CLI

anythink workflows create expense-submitted \
  --trigger Event --event EntityCreated --entity expenses \
  --enabled

anythink workflows step-add <workflow_id> send_notice \
  --name "Email finance" --action SendAnEmail --start --enabled \
  --params '{
    "to": "finance@example.com",
    "template_type": "expense_submitted",
    "payload": "{\"title\":\"{{ $anythink.trigger.data.title }}\",\"amount\":\"{{ $anythink.trigger.data.amount }}\"}"
  }'

payload is a string that holds JSON, so its inner quotes are escaped. Replace <workflow_id> with the id workflows create prints.

With an AI assistant (MCP)

Create a workflow called expense-submitted that runs when a record is created in expenses. Add a Send An Email step with the key send_notice that emails finance@example.com using the expense_submitted template, with title and amount filled from the record. Make it the start step and enable the workflow.

The assistant runs workflows create and workflows step-add through the cli tool.

The approval itself is the kanban board from step 2: dragging a card to approved updates status. To tell the submitter, add a second workflow on Entity Updated with a filter of status changed to approved, and the same email step. Test a workflow with Test Workflow in the editor, and read what each step did in its job history. Anythink doesn't retry a failed step, so check Test workflows and read job history if an email doesn't arrive. Triggers, filters, schedules and every step type are covered in How workflows work.

In the Anythink dashboard

1. Open Settings, then Organisation Settings.
2. In General Settings, set the Name and Description.
3. In the Branding card, upload a Square Logo and a Standard Logo. Each file can be up to 5 MB. The standard logo shows when the sidebar is open and the square logo when it's collapsed.
4. Under Theme, choose a Primary Color, a Gray Color and a Corner Radius. The dashboard previews each choice as you click it.
5. Select Save Changes.

With the CLI

Branding is managed in the Anythink dashboard. Use the steps above. You can choose a name and description when you create a new project:

anythink projects create "Acme Ops" --description "Acme's internal tools" --region lon1

With an AI assistant (MCP)

Branding is managed in the Anythink dashboard, and your assistant can walk you through the steps above.

The project name also appears in the emails your project sends, and your primary colour is used for their heading and button. Dark mode is a personal choice that each person makes for themselves. See Customise the Anythink dashboard.