AI interprets intent.
The model turns the request and story into a structured plan.
Security by design
AI creation.
Governed by design.
Storycards is built around explicit boundaries. AI interprets intent; Storycards authorizes and executes typed actions, validates every project, and keeps application code, credentials, access controls and publishing decisions outside the model’s reach.
Governed creation path
Enforced by Storycards
User prompt“Turn this data into an animated ranking.”
✦AIInterprets intent
No direct access to code, credentials or runtime
01
Typed actionUpdate chart
02
Product rulesAllowlisted fields
03
ValidationValid VisualSpec
✓
OutputEditable visual
JavaScriptHTML injectionCustom runtime codeOutside the AI boundary
Storycards governs execution.
Product services authorize typed actions and enforce schema and access rules.
Publishing remains a separate control.
Every release requires a dedicated preflight and an explicit user action.
Governed AI
AI works within the product model — not around it.
The model has no code editor, browser runtime or direct path to project state. It selects from defined capabilities; Storycards resolves the authorized target, constrains values and validates the resulting project.
- ✓Defined, allowlisted operations cover charts, data, copy, design, motion and layout.
- ✓Authorization stays with Storycards services for every AI-requested action.
- ✓Schema validation after every change keeps projects within the supported renderer model.
- ✓High-impact changes remain deliberate and can be presented for review before they are applied.
- ✓No arbitrary execution path exists for AI-authored scripts, markup, styles or plug-ins.
Conceptual actionNot executable code
IntentChange visual type
TargetSelected chart
Allowed valueRanking
✓
Schema gateValid and supported
ResultStructured project update
Connected data
Keep source access private. Publish only what readers need.
Storycards resolves supported live connections server-side. Readers receive a sanitized snapshot through the delivery layer — not credentials, protected endpoints or direct access to your Sheet, API or provider.
Sheets or API
Credentials are entered only through a dedicated protected connection flow — never through project chat.
Encrypted secret
Fetch and normalize
The connection is resolved server-side and normalized into the project’s structured data model.
Sanitized snapshot
Visual for readers
The public artifact excludes tokens, protected addresses and provider request details.
Isolated from content
Connection secrets are stored separately and are never copied into chat, datasets or published versions.
Public evidence, private context
Publisher citations can travel with research results while unsafe URLs and private metadata are excluded from publication.
No provider call from the reader
Published viewers read Storycards delivery snapshots rather than calling the underlying data provider.
Access governance
Access follows roles, workspaces and explicit user actions.
Storycards services enforce what each person can view, change, manage and publish. AI never expands a user’s permissions. Private projects remain outside publishing, embeds and public analytics until an authorized user moves them into an eligible shared workspace.
Explore team featuresTeam accessRole-based
G
OwnerOrganization administration and access
Full accessM
ManagerTeam administration and production
ManageE
EditorCreate and update team projects
EditV
ViewerReview without edit rights
Read onlyPrivate
Private working spacePublishing, embeds and public analytics unavailable
Identity and platform
Enterprise security beyond the AI boundary.
Storycards pairs product-level controls with identity, encryption and infrastructure protections designed for governed team use.
MFA and session control
Organizations can require authenticator-based MFA. Members can review session security, and administrators can initiate controlled recovery.
Protected in transit and at rest
TLS protects data in transit. Customer data and connection secrets are encrypted at rest and remain separate from public delivery.
Enterprise single sign-on
SAML and OIDC SSO centralize authentication so organizations can align access with their identity provider.
Managed traffic protection
Rate limiting and application-layer filtering help protect control and delivery traffic from abuse and common web attacks.
Monitoring and response
Security-relevant infrastructure and application signals support detection, investigation and coordinated response.
Ownership and deletion controls
Organizations retain control of their content and can remove projects, connections and published outputs through governed workflows.
Authoring and delivery
Your public visual is an output — not an opening into your workspace.
Storycards separates the control environment used for authoring from the delivery path used by readers. Publishing creates a validated, versioned artifact; editor state, connection credentials and management capabilities stay on the control side.
Run a publish preflight
Storycards validates the VisualSpec, data and workspace eligibility before release. Invalid output fails closed.
Create a discrete version
Each release is a distinct artifact, so teams can update, unpublish or restore deliberately.
Control embed destinations
Teams can manage embed availability and allowed domains without exposing the authoring application.
Serve through isolated delivery
Readers load a dedicated delivery bundle and CDN assets — never the editor or its management code.
Publication preflightPassed
Visual schemaValid
Credential exposureNone
Reader access to providerNone
Security controls
Defense in depth across the product workflow.
Constrained execution
The model works through defined product capabilities, with no path to author arbitrary JavaScript, HTML or CSS.
Isolated credentials
Connection secrets are encrypted, stored separately from content and resolved server-side.
Server-side authorization
Role and workspace boundaries apply at the API across AI, data and publishing actions.
Request integrity
Authenticated actions use protected session cookies, origin checks and CSRF validation.
Restricted content
Security headers constrain content types, framing and the resources published pages may load.
Versioned change
Project history and published versions make review, restore and rollback part of the workflow.
Security review
Clear answers for security and procurement teams.
Need to review the architecture or controls in more detail?
Contact StorycardsCan Storycards AI write or execute application code?
The model has no Storycards capability to edit application code or introduce executable JavaScript, HTML or CSS. It can request only typed actions within the product model.
Can AI bypass a user’s permissions?
No. AI-requested actions use the same server-side authorization as actions initiated elsewhere in the product. Read-only access remains read-only.
Which identity controls does Storycards support?
Storycards supports authenticator-based MFA, role-based access controls and enterprise SSO through SAML or OIDC.
How is data protected?
Data is encrypted in transit using TLS and encrypted at rest. Connection secrets are isolated from project content and public delivery.
Where should credentials be entered?
Only in the dedicated protected connection flow. API keys and other secrets should never be included in project chat, datasets or published content.
Does a published visual connect directly to its source?
No. For supported live connections, Storycards refreshes data server-side and publishes a sanitized snapshot. Readers receive neither provider credentials nor direct source access.
Can a Private project be published?
No. Private projects cannot be published or embedded. An authorized user must first move the project into an eligible team workspace.
What happens if a visual is invalid at publish time?
Publishing is blocked. The viewer also validates the delivered package and fails closed rather than rendering an unsupported project.
Built for governed publishing
Move at AI speed.
Keep enterprise control.
Review the architecture, controls and publishing model with our team.