Skip to content

First steps for users

OPAQUE lets you build AI workflows and analyze sensitive data—all without exposing raw data. Every action happens inside a confidential, policy-enforced workspace designed to protect data and enforce access rules by default.

Before you begin, you’ll need to sign in and optionally contribute data. Your experience will vary slightly depending on the type of workspace you’ve joined:

  • Agentic AI workspaces enable secure, policy-governed autonomous workflows.
  • Analytics & ML workspaces focus on querying and analyzing data securely.

To get started, follow these steps:

  1. Sign in to OPAQUE.

    To activate your account, sign in to OPAQUE using your organization’s SSO credentials.

    When you land on the Workspaces page, you’ll see all workspaces you’ve been added to.

    Click a workspace to open it and begin working.

    Info

    Access and execution in OPAQUE are enforced in real time based on workspace policies.

    • If you're removed from a workspace, your access is revoked immediately.
    • Running or pending workloads may be affected if underlying policies or permissions change.

    This ensures that data access and computation always reflect the latest approved policies.

  2. (Optional; Analytics & ML workspaces only) Add datasets.

    Some users may be asked to contribute data to a workspace. If that applies to you:

    • Select Datasets in the left-hand nav.
    • Click Add dataset and follow the steps to upload your schema, data, and rules.
    • Configure who can access the data and whether test data should be generated.

    For full instructions, see the Data section.

    Info

    All datasets in OPAQUE are governed by strict access controls—no other user or job can read your data unless explicitly allowed by policy.

  3. Start working in your workspace.

    What you can do inside a workspace depends on its type:

    In Agentic AI workspaces, you can:

    • Build and trigger LLM-powered workflows (e.g., summarization, routing, decision support) over sensitive data or documents.
    • Configure and launch agents that act on real-time inputs while respecting cryptographic policies.
    • View audit trails of every workflow-related decision and data access action.

    Learn about workflows.

    In Analytics & ML workspaces, you can:

    • Run notebook jobs using Python (PySpark) or SQL to analyze sensitive data without exposing it. Get started with notebook jobs.
    • Use no-code job templates to run prewritten queries or transformations—no coding required. Get started with no-code jobs.
    • Preview with synthetic data before running on real data.
    • Submit jobs for approval if policies require it, then review results within the secure workspace.

    Info

    Whether you're running queries or building autonomous agents, all actions in OPAQUE are securely executed, logged, and governed according to workspace policies.

Next steps