Explore queries Postgres, MySQL and ClickHouse.
Explore is BI without a second deployment. Write SQL against a database Connection your company already configured, watch the rows come back in the preview, and save the query as a named Chart with one of six visualizations. Pin that Chart to a 12-column Dashboard and "what was MRR last month" costs one click, for a Member or an AI Employee.
Six chart types render as SVG.
BI on the database Connections you already have: SQL saved as Charts, pinned to a 12-column Dashboard, runnable by any AI Employee.
Postgres, MySQL, and ClickHouse Connections you already have
Charts run against the Integration Connections you already configured, so there are no separate BI credentials to hold. Configs stay encrypted and are decrypted per run, with a fresh client for every query.
One save turns the SQL into a Chart
Write SQL in the editor with inline errors and a live result preview, then save it as a named Chart. Every run stops at a 30-second timeout and a 5,000-row cap.
Six visualizations render as inline SVG
Table, scalar, bar, line, area, and pie, each configured in a side panel and previewed live against your current result set. Everything renders as inline SVG, with no chart-library dependency.
Pin Charts to a 12-column drag-and-drop grid
Pin Charts as cards on a drag-and-drop 12-column grid with per-card resize. A Chart's title can be overridden per context: MRR on the finance board, Revenue (MTD) on the home one.
Every AI Employee starts at read on a Chart
Charts and Dashboards default to read for every AI Employee, authors get write on what they create, and humans grant, revoke, or promote from the Share menu.
A SELECT-only database user is the actual guard
The executor doesn't pretend to enforce read-only. Connect a SELECT-only database role, as the docs tell you to, and even a write-granted employee can't UPDATE your production data.
An employee meets the same 30-second cap.
AI Employees use Explore through ten built-in tools — list, get, and run Charts; create and update them; assemble Dashboards — under the same execution envelope humans get.
Find, don't improvise
Asked for a number, an employee lists the company's Charts, finds the right one, and runs it — reusing the SQL a human already blessed instead of guessing at schema.
Author new analytics
With write grants an employee can create Charts and Dashboards the team sees — the same way it already authors Notes and Base records.
Same limits as you
Every AI query goes through the same executor: 30-second wall-clock timeout, 5,000-row cap, credentials decrypted per run and never exposed to the model.
Readers ask 5 questions about Explore.
Which databases does Explore support?
Postgres, MySQL, and ClickHouse today. Snowflake, BigQuery, and Redshift connectors are on the roadmap.
Do I need to set up separate credentials for BI?
No. Explore reuses your existing Integration Connections — set one up under Settings → Integrations and it appears in Explore's Connection picker. Credentials stay encrypted on the Connection row and are decrypted per run.
Can AI Employees run and build charts?
Yes. Every employee defaults to read access on every Chart and Dashboard — list, get, run — and gets write on ones it authors. Humans grant or revoke per Chart or Dashboard from the Share menu.
What are the query limits?
Every execution — ad-hoc from the editor, a saved Chart, or an AI Employee's run — goes through the same executor with a 30-second wall-clock timeout and a 5,000-row cap; larger result sets are truncated server-side.
Is Explore read-only against my database?
Read-only enforcement is deliberately not baked into the executor. Connect with a separate SELECT-only database user, and no query — human or AI — can write to your production data.
4 more products share one database.
AI Employees
Teammates that carry a markdown Soul and work Routines on a schedule you pick, with every Run transcribed and costed.
Workspace
Channels, DMs, and 25 MB file uploads where AI Employees are members you can @mention.
Tasks
A task manager where any todo goes to a human or an AI Employee, and AI work stops at in_review until you sign it off.
Bases
Multi-table workspaces with 11 field types, saved views, comments, and attachments, plus 21 built-in tools for a granted AI Employee.
Tomorrow one thing is finished before 09:30.
Install Genosyn, register an AI Model, write one role, and put it on a schedule. The company grows from there.