Directly
A server the cloud can reach
Point it at the server with a read-only account. The setup shows the exact GRANT script for your database, and a test reports what it found before anything is saved.
Connect a databaseBetaFree accounts and the desktop app are available now. Paid plans, with reports on the web, sharing and live connections, launch once enough people ask for them. Express your interest
Connect your data Plus & Ultimate · add-on
Coming soon Live connections from the website come with the full release. The desktop app already reads SQL Server, PostgreSQL, MySQL, Oracle and ODBC databases on your network, at no charge. The full release launches once enough people ask for it. Express your interest
Connect SQL Server, PostgreSQL, MySQL or Oracle and refresh a dataset from it as often as every 15 minutes. Reports read the refreshed copy, so opening or sharing one never touches your server — unless you make a dataset live, and then it reads the database the moment the report opens.
The Data sources add-on is $30 a month per workspace, on Plus and Ultimate.
Coming soon: Adding a SQL Server data source, testing it, making a dataset and setting a 15-minute refresh
Three ways in
Directly
Point it at the server with a read-only account. The setup shows the exact GRANT script for your database, and a test reports what it found before anything is saved.
Connect a databaseFrom the cloud
Azure SQL and Azure Database for PostgreSQL or MySQL, Amazon RDS and Aurora, and Google Cloud SQL can trust your workspace’s own identity for that one data source. There is no long-lived secret anywhere.
Cloud sign-inThrough a gateway
A small service on a Windows, Linux or macOS machine you choose. It connects outward only, needs no inbound firewall rule, and its credentials never leave the machine.
Install a gatewayLinked data
Customers and orders, projects and tasks, budget and actuals: link two datasets on a shared column, from a database or a file, and the report works across both.
Refresh
Choose how often — when asked, every 15 minutes to 24 hours, daily or weekly — and whether each run replaces the rows or adds a dated snapshot. Every run is listed with who or which schedule started it, what it loaded and how long it took.
Coming soon: A data source’s run history, showing start times, rows loaded, durations and a failed run explained in plain words
More than a table
Stored procedures
A dataset can come from a procedure as easily as a table. Its parameters become the dataset’s inputs, set once or filled in when the report opens.
Query texts
Whoever manages a data source can write the query a dataset reads. It is scanned for anything that could change data, and an admin reviews it before the first run.
Live datasets
For the numbers that can’t wait for a schedule: a live dataset reads the data source each time the report opens, can refresh itself while the page is up, and can take the reader’s own identity as an input.
Write-back
Give a dataset its key fields and a write action — an update, insert or delete through a table, a stored procedure or a query you write — and an edit in the grid is written back where the row lives.
See exactly which rows a write would touch before anything is changed.
Who changed what, when, and what the database said back.
Through a gateway, writes are approved on the machine inside your network, and a data source can be sealed so nothing writes to it at all.
How write-back worksSpreadsheets too
Not everything lives in a database. Drop another file onto an open report and it becomes another dataset, ready to link to the first. Save the report, and next month’s file drops straight in.
Kept safe
Each workspace gets its own vault, identity and key service, so credentials are never shared with another workspace.
Encrypt stored credentials under a key in your Azure Key Vault, AWS KMS or Google Cloud KMS. Disable it, and refreshes stop from the next run on.
Before the first data source, every member must use an authenticator app, and every change to a data source asks for a code.
A dedicated account with SELECT on only the tables you choose. Reports read the stored copy; only a live dataset or a write action you set up reaches the database.
Reach an Azure database over Private Link with no public endpoint, or let a gateway sign in to SQL Server as the Windows account it runs under. Each data source also says whether public reports may read it.
Coming soon: Pairing an on-premises gateway, with its pairing code and the prompt to approve the machine
Data sources add-on
The workspace owner adds it on Plan & billing, the same way as extra users. Every person in the workspace can use the data sources they’re given.
SQL Server, PostgreSQL, MySQL / MariaDB and Oracle — on your own servers, or in Azure, AWS or Google Cloud.
Only refreshes reach it, on the schedule you choose, with a read-only account. Reports and shared links read the stored copy, so a busy dashboard puts no load on your server.
A copy by default, refreshed as often as every 15 minutes. That is what keeps your server out of every page view, and what lets a public link work without anyone reaching your network. A dataset you mark as live reads the database when the report opens instead, for the few numbers that can’t wait.
Yes. Give the dataset its key fields and a write action, and edits made in the grid are written to the table, procedure or query you chose, after a dry run and with a history of every write.
Yes, through a gateway installed on a machine inside your network. It makes one outbound connection, so there is no inbound rule to open.
In the workspace’s own vault, or on the gateway machine for a gateway data source. Cloud sign-in needs none at all.
The workspace owner and admins, with a code from their authenticator app for every change. Everyone else uses the data sources they are given.