> ## Documentation Index > Fetch the complete documentation index at: https://docs.hiveku.com/llms.txt > Use this file to discover all available pages before exploring further. # Your Data > Tables of your own data, filled by workflows, people and AI agents, that dashboard widgets read like any other data **Your data** holds the account's own tables: numbers from tools outside Hiveku (a point-of-sale system, a warehouse, a spreadsheet), or anything your team keeps by hand. Dashboard widgets read them like any other Hiveku data, and update within seconds of a change. ``` Settings > Your data ``` You can also reach it from the bottom of the dashboard list. Each table is a spreadsheet-like grid with typed columns. ## Kinds of table | Kind | What it is for | | - | - | | **Running log** | Each new row is something that happened: a sale, a sign-up, a ticket from another tool. Every row has a **When it happened** time, and widgets count rows and add up numbers over the dashboard's dates. | | **Latest values** | One row per key, updated in place: stock per warehouse, a bank balance, a queue length. Each row has a **Key** that names it, such as an item or a place. Widgets show the latest values right now. | | **Tally** | A count you add to, one at a time: walk-ins, calls about a promotion. Every add is kept, with who added it. See [Tallies](#tallies). | ## Make a table * **New table**: give it a name and a kind, then add its columns. * **Import CSV**: choose a file (up to 10 MB and 10,000 rows) or drop it on the window. Before anything is saved you see the table's name and kind, each column with the name and type it will get (change either, or leave a column out), and every row checked, with bad cells marked in red and why. **Import** makes the table and adds the good rows. Values are read as written and never run as formulas: a cell that starts with `=`, `+`, `-` or `@` is kept as that text. Times in a file have no time zone, so they are read in the account's time zone, and the window says which. ## Columns A table has up to 20 columns. Each has a type: | Type | Holds | | - | - | | **Text**, **Long text** | Words. Text is cut at 500 characters, Long text at 1,500. | | **Number**, **Percent** | Numbers. | | **Money** | Amounts in one currency, set for the column, so amounts are never added across currencies. | | **Date**, **Date and time** | Days, or moments in the account's time zone. | | **Yes/No** | A tick. | | **Choice**, **Multiple choice** | One or several options from a list, each with its colour. | From the table's menu you can add a column, and rename, hide or move one. Renaming changes only the label, so workflows and widgets that use the column keep working. **Changing a column's type** or **deleting a column** asks first and says what will happen. Both can be undone for 30 days from **Column changes**. ## Rows * **Add row**, or click a cell to change it in place. * **Paste rows** from a spreadsheet. Each cell is checked against its column, and nothing is saved until the bad rows are fixed or left out. * **Search**, **Filter** and **Sort** work over the whole table, not just the rows on screen. * **Delete** rows you select. Deleted rows can be restored for 30 days. * **Export** the table as CSV. A row a workflow writes says so when you open it: the workflow will replace your change on its next run. ## Fill a table from a workflow The **Send to Dashboard** workflow step saves rows from earlier steps into one of your tables. Two templates under **Dashboards** in the workflow template list set it up: * **Webhook to Dashboard**: another tool sends its data to a webhook, and each delivery is saved as a row. Two settings on the webhook come first: * Make its key with **Generate a random secret** under **Authentication**. Deliveries are refused until it has one, and the tool sends the key in an `X-API-Key` header. * Set **What arrives here** to the table, so deliveries are saved as rows and not as leads. It starts on leads, and only an owner or admin can change it. * **Check an API on a Schedule**: every hour, read another tool's API and save each item it returns as a row. Keep the tool's key in the workflow's environment and refer to it as `{{env.API_TOKEN}}`; it is never written into a table. A workflow made from either template starts switched off. Choose the table in the Send to Dashboard step, test it, then switch it on. A step can write one row per item of a list, up to 500 rows a run. A row with an id of its own (an order number, say) is saved once, even if the tool sends it again. Every run counts toward the plan's workflow runs, like any other run. ## Tallies A tally counts something Hiveku can't see for itself, such as patients who walk in without a booking. Before making one, check whether Hiveku already counts it: bookings, calls, emails and deals are counted from the real records. A tally shows its totals today, this month and all time, an **Add 1** button for the people it lets add, and its **Activity**: every add, newest first, with who made it and their note. **Tally settings** (owners and admins): * **What it counts**: a few words, such as "Walk-ins". AI agents read these words, so they can't hold a link, an email address or a phone number. * **When to add**: in your own words, for the people who add. AI agents never read it. * **People who can add**: everyone who can see it, or only people who can change Your data. * **Workflows can add**, with a Send to Dashboard step. * **AI agents acting for a person can add**: an AI app on that person's own Hiveku key, or the chief of staff working for them. Each add points to the record it counted, such as the booking or the call, so the same record never counts twice. * **Most one add counts**, **Most an hour** and **Most a day**: past the most an hour or a day, adds are refused until the hour or day moves on, and the account's inbox says so. A wrong add can be taken back: **Take it back** asks why, and the add stays in the activity, struck out, with who took it back and the reason. People who can delete Your data can also **Start from zero** (every add is kept, and Undo counts them again). A tally can also sit on a dashboard as a widget with its own **Add 1** button. ## Who can do what | Role | Can | | - | - | | **Everyone** | See the tables, and widgets on them. (A table an AI app made for managers only shows only to people who can change Your data.) | | **Owners and admins** (or a role with **Your data: Create / Edit**) | Make tables; add, change, delete and restore rows; add, rename and hide columns. | | **Owners and admins** (or a role with **Your data: Delete**) | Delete and restore tables, delete columns, change a column's type, start a tally from zero. | Deleted tables can be restored for 30 days from **Deleted tables**. ## Limits | Limit | Amount | | - | - | | Tables (running logs and latest values) | 25 per account | | Tallies | 50 per account | | Columns | 20 per table | | Rows | 100,000 per table, and at most 10,000 added an hour; running-log rows that happened more than 400 days ago are dropped | | Rows per workflow run or agent call | 500 | | One row | 2 KB | This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.