ADMIN GUIDE

Configure the workspace for your team.

Use this guide after dig data is running. Each topic explains what the setting does, who can use it, how to configure it, and what to check when something goes wrong.

Admin topicsFirst administratorTeam membersData sourcesDictionaryAccessRelationshipsAI settingsAudit logSecurity and retention

Create the first administrator

What it is for: Create the account that controls the workspace.

Who can use it: The first person opening a new workspace; later changes require an administrator.

  1. Open the dig data URL while the workspace has no account.
  2. Enter your name, work email, and a password of at least 12 characters.
  3. Select Create administrator.
  4. Sign in, then configure AI settings, Data sources, and Team members.

The first account is an administrator. Store its credentials securely and avoid sharing the account with the team.

Create employee accounts

What it is for: Give teammates their own sign-in instead of sharing credentials.

Who can use it: Administrators.

  1. Open Admin → Team members.
  2. Select + Add employee.
  3. Enter the employee's name, work email, and a temporary password of at least 12 characters.
  4. Select Add employee and send the temporary password privately.
  5. Ask the employee to change it on their first sign-in.

If it does not work: Check that the email is valid and that the temporary password meets the minimum length.

Create a database source

What it is for: Make approved PostgreSQL, MySQL, MariaDB, or MongoDB data available to the workspace.

Who can use it: Administrators.

  1. Open Admin → Data sources.
  2. Select + Add source.
  3. Enter a clear source name and choose the database type.
  4. For MongoDB, enter the database name when prompted.
  5. Enter the connection URI and select Test and add source.

Security: Use a dedicated read-only database account limited to approved schemas, tables, or collections. Employees can ask about everything that account can read.

If it does not work: Confirm the container can reach the database, use host.docker.internal for a database on the Docker host, and check the URI, credentials, firewall, and TLS settings.

Create and manage Dictionary definitions

What it is for: Keep the team's business language consistent in answers and dashboards.

Who can use it: Everyone can create definitions. The creator and administrators can edit them.

  1. Open Dictionary from the workspace navigation.
  2. Select + Add definition.
  3. Enter the term and a precise meaning, then save it.
  4. To update it, find the definition and select its edit action.
  5. Use @term in a question to make the intended definition explicit.

Each edit creates a new version. Existing dashboard insights retain the version used during preview until they are previewed and saved again.

Grant source access

What it is for: Decide which employees can ask questions about each source.

Who can use it: Administrators.

  1. Open Admin → Access.
  2. Select + Grant access.
  3. Choose an active employee and a data source.
  4. Select Grant access.
  5. Remove the grant from the same list when access should end.

Security: A grant does not narrow database permissions. The employee can ask about anything the source credential can read, so keep both grants and database privileges narrow.

Configure source relationships

What it is for: Allow cross-database questions to match fields between two assigned sources.

Who can use it: Administrators.

  1. Open Admin → Relationships.
  2. Select + Add relationship.
  3. Choose the left source, table or collection, and field.
  4. Choose the right source, table or collection, and matching field.
  5. Use Refresh schemas if a table or field is missing, then select Save relationship.

If it does not work: Confirm both sources are connected and that their credentials can read the selected objects.

Configure the AI provider

What it is for: Choose the provider and model used to answer questions.

Who can use it: Administrators.

  1. Open Admin → AI settings.
  2. Select Configure.
  3. Choose Anthropic or OpenAI.
  4. Enter the provider API key and model ID.
  5. Select Save AI settings, then test it with a safe question.

Security: Keep the key private. For database questions, relevant schema details, the question, and bounded query results may be sent to the selected provider.

If it does not work: Confirm the provider, key, model ID, and outbound network access from the dig data host.

Review the audit log

What it is for: Inspect questions, answers, generated queries, and failures across the workspace.

Who can use it: Administrators.

  1. Open Audit from the workspace navigation.
  2. Filter by All, Successful, or Errors.
  3. Open a record to inspect its generated query when available.
  4. Select the CSV export to download the current filter or all retained records.

Audit questions, answers, and query results are retained for 90 days. Only administrators can open the audit API or export records.

Manage security and retention

What it is for: Keep the deployment, credentials, grants, and retained records under organizational control.

Who can use it: Administrators and the infrastructure owner.

  1. Keep dig data on a private company network or VPN and use HTTPS through your existing reverse proxy.
  2. Back up the dig-data volume securely; it contains accounts, connection URIs, provider keys, dashboards, and audit records.
  3. Use read-only source credentials and review grants regularly.
  4. Follow the Setup guide for consistent backups, restores, upgrades, and health checks.

Credentials and retained audit content are not encrypted at rest in this beta. Protect the host, volume backups, reverse proxy, and administrator accounts accordingly.

← Back to Documentation · Read the Setup guide →