Docs
How orbstrata works: installing and pairing the agent, connecting your AI client, the console's places, and the command line.
Start here
Your databases
- Add a databaseThe agent reads PostgreSQL, MySQL, MariaDB, SQL Server, Oracle and SQLite.
- Make a read-only roleThe agent reads every database through a role that can only read.
- Reflect a sourcestrata reflect crm reads a source's catalog, never a row:
- Tags and maskingA tag says what a column holds:
- ProposalsNobody's suggestion changes the map until a person applies it.
- Read the queries logQueries is the audit, led by the statements asked of your sources:
The console
- The consoleThe console is where a workspace is seen and governed.
- Your accountYou sign in with your address and a password of at least 10 characters.
- OverviewOverview is the first place:
- ConnectionsConnections lists the databases your agents reach:
- SourcesSources lists every database your agents serve:
- SchemaSchema shows a source's tables and columns as the last reflection read them, never a row.
- MapMap is the semantic map as people applied it, across every source.
- AgentsAgents lists the machines paired with this workspace.
- PeoplePeople lists who is in this workspace, each one's role, and the invitations waiting.
- PolicyPolicy is where an Owner decides what each role may do on each tier, and which tier a name falls in.
- SimulatorSimulator answers "what would happen if…" before anyone publishes.
- SettingsSettings holds this workspace's name, its plan, its API tokens, your account, and deleting.
Reference
- The command linestrata is the agent:
- Roles, tiers and rulesEvery tool call is decided on the agent before anything runs, by the role of the member who paired it and the workspace's rules.
- Connect claude.aiclaude.ai and Claude Desktop reach your workspace through its gateway by signing you in, so they hold no token you made.
- RunsRuns is where you write a playbook:
- Plans and billingEvery workspace is on a plan.