Interactive artifacts

Draft, publish, and share interactive artifacts directly from Glean.

Overview

Interactive artifacts are rich, dynamic pages created from canvas that are ideal for sharing stable content with a link in Glean. Under the hood, these artifacts are powered by HTML rendered in a secure sandbox.

Unlike export-first content types (for example, docs, slides, emails, and Slack posts), interactive artifacts:

• Live natively in Glean. Users can access interactive artifacts with a stable URL or optional go/link.
• Are independent from the original chat, so you can safely share the finished experience without exposing the conversation that produced it.
• Show up in Glean search, just like other key knowledge surfaces.

Enable Canvas and Glean for your organization to use interactive artifacts.

When to use interactive artifacts

Use interactive artifacts when:

• The output is dynamic, visual, or non-document-shaped (for example, exploratory views, dashboards, walkthroughs, light apps).
• You want a single, stable link that people can bookmark, or turn into a go/link, and you don't need Google Docs/Slides collaboration features.
• You want to share the output, but not the chat. For example, a polished interactive explainer generated after a long Glean session.

Use an export-first content type when:

• The canonical home should be an external editor: Docs, Slides, Email, Slack/Teams, or similar.
• You need rich, native collaboration (comments, tracked changes, presentation mode) in those tools.

Common use cases

Interactive artifacts are a good fit for:

• Interactive explainers and walkthroughs

  • Product tours, architectural diagrams, or "how this system works" pages where interaction helps tell the story.

• Dashboards

  • Interactive views that summarize recent activity, metrics, or findings for a customer or initiative, shared via a stable link.

• Lightweight internal apps and tools

  • Simple calculators, forms, or interactive visualizations generated by Glean and shared as a self-contained experience for your team.

• Research and analysis hubs

  • Interactive visualizations for Deep Research outputs or internal analyses where a static document doesn't capture the full experience.

Key capabilities

Interactive artifacts support:

• Glean-native publishing

  • Sharing an interactive artifact publishes the current Canvas HTML view into a standalone page.
  • The artifact is coupled with further chat changes. You can continue iterating in Canvas, and viewers see the latest version.

• Audience and role controls

  • Choose between private, team, or org-wide visibility.
  • Assign view vs. edit access for collaborators.

• Stable URLs and go/links

  • Glean assigns a stable URL for every interactive artifact, with an option to attach a human-friendly go/link (for example, go/your_app).

• Discovery across Glean

  • Interactive artifacts are visible in Glean search, alongside other important work outputs.

• Context jump-back (for editors)

  • Editors can jump from the published interactive artifact back to the originating Canvas/chat to see how it was created and continue iterating.

How to use interactive artifacts

Step 1: Create in Canvas

Start in Glean Canvas as you would for any content type: ask Glean to generate an HTML walkthrough, mini-app, or visualization, or paste in existing HTML and refine it with Glean.

Step 2: Preview

Canvas renders the HTML inside a secure, sandboxed frame so you can see the final interactive experience before publishing.

Step 3: Publish and share

When you're satisfied, select Share > General access > Anyone at <Organization> with the link can view to publish it.

• Glean assigns a stable URL and (optionally) a go/link.
• You can share this link in email, Slack/Teams, or docs without exposing the underlying chat.

Permissions and sharing

Interactive artifacts follow Glean's existing permissions model:

• Share within Glean only

  • Sharing an interactive artifact changes who can see it in Glean (Search, direct link). It doesn't change permissions in any external systems.

• Audience presets

  • Private to you
  • Visible to a specific team
  • Org-wide, for broad-reach artifacts like internal microsites or dashboards

• Access controls

  • Viewers can open and interact with the experience (within the sandbox).
  • Editors can update the underlying Canvas and republish updated versions as needed.

Where applicable, Glean may add light attribution (for example, a "Generated with Glean" footer or URL path) so new viewers understand that the artifact was produced in Glean and can discover Glean for their own workflows.

Security and sandboxing

Because interactive artifacts can contain dynamic HTML, CSS, and JavaScript, Glean executes them inside a locked-down sandbox designed to protect your data and session:

• Sandboxed iframe execution

  • HTML runs inside an iframe with sandbox="allow-scripts", which isolates it from the parent app and gives it a null origin.
  • The code cannot access Glean cookies, localStorage, sessionStorage, or the parent DOM.

• Strict content security policy (CSP)

  • A dedicated CSP for the sandbox blocks outbound network requests (connect-src 'none'), disallows nested frames, and restricts form submissions.
  • Images and fonts are limited to safe sources to prevent data exfiltration via URLs.

• No access to user data or credentials

  • Executed HTML cannot read Glean authentication tokens, user credentials, or private document content beyond what the artifact is explicitly rendering.

These controls are part of Glean's broader security posture for Canvas and HTML artifact execution; security teams can review the detailed architecture in the Artifact HTML Code Execution design and related security documentation.

Pricing and credit usage

Preview period

Generating interactive artifacts doesn't consume FlexCredits during the preview period. Admins will be given at least two weeks' advance notice before the preview period concludes.

Interactive artifacts are a premium feature. Generating them consumes FlexCredits.

Known limitations

Interactive artifacts are powerful but intentionally constrained to keep them safe and maintainable:

• No server-side code

  • Interactive artifacts run only client-side HTML/CSS/JS. They don't support Python, SQL, or server-side execution.

• No outbound network calls

  • The sandbox CSP blocks network requests (connect-src 'none'), so interactive artifacts cannot call external APIs or arbitrary endpoints.

• No nested iframes or top-level navigation

  • Interactive artifacts cannot embed additional frames or navigate the parent page (for example, redirecting the Glean app).

• No persistent storage for executed code

  • Interactive artifacts don't provide persistent storage for client-side state beyond what's in the artifact itself; reruns start from a clean slate.

If you need deep, collaborative editing, advanced formatting, or long-term versioning, consider exporting to Docs/Slides instead and using interactive artifacts for the interactive or exploratory parts of the workflow.

FAQs

Yes. You can continue editing in Canvas, and viewers with the shared link will always see the latest version.

No. Viewers see only the finished interactive experience. Only editors with appropriate access can jump back to the originating chat/Canvas for additional context.

They appear in Glean search, alongside other key content created via Canvas.

No. For security reasons, Glean blocks outbound network requests and nested frames in the sandbox. If you need deep integration with external systems, consider surfacing those systems directly via Glean Agents and actions instead.

See also

• Glean Canvas - Overview of drafting, iterating, and exporting written artifacts in Canvas.
• Image generation in Glean - How to generate visuals that you can embed into interactive artifacts and other outputs.