Blokay

The Explorer is where most work starts. You write a SELECT statement, run it, and convert the result to a view with one click. Blokay builds the table, keeps the filters and turns hardcoded values such as a status or a city into controls the reader can change. The query stays behind the view, so nobody downstream edits SQL to get a different slice.

Describe what you need in plain language and the agent reads your tables, writes the query, assembles the component and leaves it ready to share. It runs against your database with the permissions of whoever asks, so a read-only user can't get a write action out of it. The prompt bar also suggests starting points like pending orders by city or a sales by month chart.

Tables, columns, relations and sample rows are browsable without installing a client. You can run SQL from the same screen, and people who only query get read-only access by default. Private datasources stay hidden from users who shouldn't see them, the AI included.

Three roles cover most teams: admin, editor and user, with finer permissions set per person. Each user sees only the views assigned to them, and each datasource can be hidden from specific people. The sandbox and SSH tunnel options let you test against staging or reach a private database without exposing it to the internet.

Any view or block can run on a schedule, from every fifteen minutes to a weekly report on Monday morning. Runs show up in the history with what happened and how long it took, which is the difference between a job you trust and one you check by hand. Alerts can go out when a threshold breaks, such as three products below minimum stock.

Every view is a web component. Drop it into HTML, React, Vue, Next.js, Angular or Svelte and your customers see their data with their own session, without you building a separate back office. That covers the case where an internal tool turns out to be a customer feature.