CLARA does not transmit or store prompt text, message text, page content, clipboard text, or file contents. Certain risk checks may inspect pasted or submitted text locally in the browser to determine whether policy action is required.
For the C.L.A.R.A. Chrome Extension and Related Services
Privacy Policy
This Privacy Policy explains how the C.L.A.R.A. Chrome extension and related service process information for enterprise AI governance. Sensitive-input risk detection runs locally in the browser extension. CLARA's backend and dashboards use structured metadata rather than prompt text, AI response text, or page-content payloads.
Policy overview
Designed to give organisations AI governance without hidden monitoring.
This policy applies to the C.L.A.R.A. Chrome extension, related administrative dashboard, associated APIs, customer support, and operational services used to deliver AI governance, visibility, policy enforcement, transparency, and evidence workflows.
Sensitive-input risk detection runs locally in the browser extension. Backend classification and reporting use metadata such as tool, domain, event type, severity, and policy outcome.
C.L.A.R.A. is not designed as a covert monitoring, keystroke logging, screenshot capture, or employee surveillance product.
1. What data we collect and process
Depending on customer configuration and how the service is deployed, C.L.A.R.A. may process structured governance and account information needed to operate the extension and related service. This includes metadata such as:
Certain risk checks may locally inspect pasted text and submitted input values in the browser extension to determine whether policy action is required. That local inspection is used for in-browser risk detection; only structured metadata is sent to the backend.
- Account and tenant information: organisation name, tenant identifiers, administrator-set roles or groups, subscription and configuration data.
- Extension and device metadata: extension version, browser type, environment identifiers, and technical diagnostics needed to secure and maintain the service.
- Tool, domain, and event metadata: the identity of an AI tool or destination domain, timestamps, policy outcome labels such as allow, warn, or block, severity categories, reason codes, and audit evidence describing the governance event.
- Administrative records: policy settings, approved or blocked tool lists, configuration changes, access logs, and support records associated with customer administration of the product.
2. What data CLARA does not transmit or store by default
CLARA does not transmit or store prompt text, message text, page content, clipboard text, or file contents. Certain risk checks may inspect pasted or submitted text locally in the browser to determine whether policy action is required. Only structured metadata is sent to the backend.
- Prompt text or message text is not transmitted to or stored in CLARA's backend or dashboards.
- AI response text is not transmitted to or stored in CLARA's backend or dashboards.
- Page content payloads are not transmitted to or stored in CLARA's backend or dashboards.
- File contents are not read or stored by the current extension logic.
- Screenshots, screen recordings, or hidden visual capture.
- Keystrokes, keylogging data, microphone input, camera data, or hidden monitoring feeds.
If a customer requests additional logging or custom workflows beyond the default metadata-only model, that processing should be documented separately in the applicable agreement, deployment documentation, or customer notice.
3. How we use data
We use structured metadata and related service information to provide and improve the C.L.A.R.A. service, including to apply governance policy, present dashboards, provide employee transparency records, investigate incidents, maintain security, prevent abuse, respond to support requests, and comply with contractual or legal obligations.
Sensitive-input risk detection runs locally in the browser extension. Backend classification and reporting use metadata such as tool, domain, event type, severity, and policy outcome. CLARA's backend and dashboards do not receive or store prompt text or AI response text.
We may also generate aggregated or de-identified statistics that help us understand system performance, product reliability, or overall governance trends, provided that those outputs do not include prompt text, AI response text, page content, clipboard text, or file contents.
4. Sharing and disclosures
If enabled by the customer, CLARA may send metadata-only governance events to customer-configured integration endpoints, such as webhook destinations selected by the customer. These payloads are designed not to include prompt text, response text, page content, clipboard text, or file contents.
We may share relevant metadata and service information only in the following circumstances:
- With the customer organisation: authorised administrators and designated personnel may access governance metadata and policy outcomes associated with their tenant.
- With customer-configured integration endpoints: metadata-only governance events may be sent to integrations or webhook destinations selected and enabled by the customer.
- With service providers: infrastructure, hosting, security, support, and other subprocessors that help us operate the service under confidentiality and data protection obligations.
- For legal and safety reasons: where required by law, regulation, court order, or to protect rights, security, or the integrity of the service.
- In a business transaction: in connection with a merger, acquisition, restructuring, or financing, subject to applicable confidentiality and legal protections.
5. Retention
CLARA includes retention settings for structured metadata event records and admin audit records. These settings are currently a pilot-stage foundation. Automated retention enforcement and cleanup workflows are not yet fully implemented in production.
As a result, actual retention may vary based on deployment, storage practices, operational needs, customer configuration, and applicable legal obligations.
6. Security
We use technical and organisational measures intended to help protect structured metadata and related service records against unauthorised access, disclosure, alteration, or destruction.
Because sensitive-input risk detection runs locally in the browser extension and backend systems use structured metadata rather than prompt or response text, CLARA reduces exposure by limiting the amount of readable content sent to backend systems.
7. Employee transparency
C.L.A.R.A. is intended to support transparent enterprise governance, not covert workforce monitoring. The service is designed so organisations can explain what metadata is recorded, what local risk checks occur in the browser, and what content is not transmitted or stored by the backend.
Where implemented in a customer deployment, employees may be given a self-view or comparable record showing governance metadata, policy alerts, and related notices tied to their usage context. Customers are responsible for providing workplace notices, internal policies, and legal disclosures appropriate for their deployment.
8. Legal basis and compliance-friendly framework
Depending on deployment, contract structure, and applicable law, personal data processed through CLARA may be handled under contractual necessity, legitimate interests in enterprise security and AI governance, legal obligations, consent, or other bases permitted by applicable law.
In many enterprise deployments, the customer organisation determines workplace use, policy settings, and employee notices, and CLARA may operate as a processor or service provider on the customer's behalf. Individuals should normally direct workplace privacy questions or rights requests to their organisation first.
9. International transfers and rights
Depending on where CLARA is deployed and hosted, service data may be processed across jurisdictions. Applicable transfer safeguards and individual rights will depend on the relevant deployment, contract, and law.
Enterprise users should normally direct workplace-related privacy questions or requests to their organisation first. CLARA may assist customer organisations as required by contract or applicable law.
10. Contact Information and Updates
We may update this Privacy Policy from time to time to reflect changes in the service, applicable law, or operational practices. When we do, we will update the effective date above and post the revised version on this page.
For privacy questions, data protection inquiries, or requests related to this Privacy Policy, please contact:
hello@thisclara.com
If required, we may also provide the relevant legal entity name, postal address, and data protection representative details for the applicable CLARA service operator.