Privacy Policy

Last updated: April 11, 2026

This Privacy Policy explains how SemSwitch, Inc. collects, uses, stores, and shares information in connection with soundOS.

For purposes of this policy, soundOS includes the public website at soundos.app, the authenticated dashboard, the current Windows desktop application, and the hosted APIs and service components that support those experiences.

soundOS is an agent orchestration product with voice, session, and workspace features. Some parts of the product run locally on your device, and some parts run in hosted services behind the soundOS website and application surfaces.

1. Information soundOS collects

Account and identity information

If you sign in to soundOS, we may collect and process account-related information such as:

  • name or display name
  • email address
  • avatar or profile image
  • account and organization identifiers
  • authentication and session records

The current sign-in architecture uses Clerk-backed authentication flows for soundOS account access.

Workspace and configuration data

soundOS may store and process workspace and product configuration data such as:

  • account settings
  • interface preferences
  • voice preferences
  • topology or orchestration state
  • agent profiles
  • squads, session metadata, and related workspace configuration

Some of this data is cloud-backed so it can be retrieved across signed-in product surfaces.

Content you provide or generate

Depending on how you use soundOS, we may process:

  • prompts
  • commands
  • text inputs
  • voice inputs
  • transcripts
  • files, artifacts, or attachments you choose to submit through supported workflows
  • generated outputs
  • session history
  • approvals, receipts, and evidence-related records tied to workflow execution

Voice and speech data

If you use voice features, soundOS may process:

  • microphone audio
  • partial and final transcripts
  • text-to-speech requests and playback artifacts
  • session identifiers and timing metadata
  • approval, execution, and evidence records tied to a voice-driven workflow
  • receipt and lineage records for turns or actions

Voice features are optional, but they necessarily involve processing speech-related data in order to function.

Technical, device, and operational data

We may collect service and device metadata needed to run and secure soundOS, including:

  • request and response metadata
  • timestamps
  • technical logs
  • session identifiers
  • workspace or user scope identifiers
  • error and diagnostic information
  • event stream or runtime state needed to keep the interface in sync

Desktop-local data

The desktop app stores some information locally on your device. Based on the current soundOS codebase, that includes:

  • non-secret profile cache data in application data directories
  • local device settings and runtime preferences
  • active or recent terminal session state needed for the desktop experience
  • provider API keys and certain refresh/session material in the operating system credential manager or keychain

In the current desktop implementation, sensitive provider API keys are stored in the OS credential manager rather than written into your workspace folder.

2. How soundOS uses information

We use information collected through soundOS to:

  • authenticate users and manage sessions
  • provide the website, dashboard, desktop app, and hosted services
  • sync account settings and workspace state across supported surfaces
  • power orchestration, session history, artifacts, approvals, and voice workflows
  • launch and manage desktop-local features where applicable
  • troubleshoot errors and maintain product reliability
  • detect abuse, protect the service, and enforce our Terms and EULA
  • comply with legal obligations
  • communicate with users about account, product, billing, or security matters

We do not need to use every category of information for every feature. What is processed depends on the parts of soundOS you use.

3. How information may be shared

Service providers and infrastructure

soundOS relies on third-party infrastructure and service providers to operate the product. Based on the current codebase, this includes services used for:

  • authentication and account access
  • hosted API delivery and storage
  • session and state handling
  • artifact and receipt storage
  • speech processing and voice services
  • model or tool-provider connectivity where features require it

The current architecture includes Cloudflare-hosted service components, Clerk-backed authentication, and a voice path that uses Deepgram through Cloudflare AI Gateway for current speech processing flows.

Model and provider calls you invoke

If you connect external model providers, bring your own provider keys, or invoke provider-backed features through soundOS, your prompts, commands, content, transcripts, or related metadata may be sent to the relevant provider as needed to fulfill your request.

Legal and safety reasons

We may disclose information where reasonably necessary to:

  • comply with law, regulation, legal process, or a valid governmental request
  • enforce our agreements
  • detect, prevent, or respond to fraud, abuse, or security issues
  • protect the rights, property, or safety of SemSwitch, Inc., soundOS, our users, or others

Corporate transactions

If SemSwitch, Inc. is involved in a merger, acquisition, financing, reorganization, or sale of assets, information related to soundOS may be disclosed or transferred as part of that transaction, subject to applicable law.

4. Cookies, local storage, and device permissions

Browser sessions

The authenticated dashboard uses browser session mechanisms necessary to keep you signed in and to secure requests. In the current implementation, after authentication soundOS exchanges the sign-in token for a secure HttpOnly application session cookie.

Authentication providers used by soundOS may also use cookies or similar browser mechanisms needed to complete sign-in and session handling.

Desktop-local storage

The Windows desktop app stores certain local data under operating-system-managed app data locations. The current app also uses the OS credential manager for sensitive items such as saved provider API keys and certain refresh/session material.

Microphone access

If you use voice features, soundOS will request microphone access from your device or browser. Voice features will not function without that permission.

5. Retention

soundOS retention depends on the type of data and the feature involved.

Examples from the current implementation include:

  • browser application sessions are currently configured with a maximum lifetime of up to 72 hours unless revoked or expired sooner
  • desktop-local auth cache and profile data may remain on the device until sign-out, uninstall, replacement, or local cleanup
  • provider API keys saved in the desktop app may remain in the OS credential manager until you remove them
  • cloud-backed settings and workspace state may remain until changed, deleted, or no longer needed for the service
  • voice artifacts may be time-scoped in some flows, while related receipts, history, or evidence records may persist longer where needed for product function, auditability, support, or security

We may also retain logs, security records, and operational metadata for as long as reasonably necessary to operate, secure, and support soundOS, resolve disputes, and comply with legal obligations.

6. Security

soundOS uses a combination of local-device protections and hosted service controls. Based on the current codebase, those measures include secure session handling, workspace-scoped access controls, OS keychain storage for certain desktop secrets, signed internal service flows for sensitive voice and artifact operations, and integrity checks in the current Windows desktop distribution path.

No internet-connected service can promise perfect security. You should use soundOS with appropriate judgment, especially when handling sensitive data, local execution features, voice features, or external provider connections.

7. Your choices

Depending on how you use soundOS, you may be able to:

  • sign out and revoke active sessions
  • remove locally stored provider API keys from the desktop app
  • uninstall the desktop app and clear local data from your device
  • choose whether to use voice features
  • adjust account-level settings and preferences

If you have a privacy request or question related to soundOS, contact:

SemSwitch, Inc.
8 The Green, Suite B
Dover, DE 19901
legal@semswitch.com

8. Changes to this policy

We may update this Privacy Policy from time to time. If we make material changes, we may update the date above and, where appropriate, provide additional notice.

Your continued use of soundOS after an updated Privacy Policy takes effect means the updated policy will apply to your use of the service.

9. Contact

If you have questions about this Privacy Policy or how soundOS handles information, contact:

SemSwitch, Inc.
8 The Green, Suite B
Dover, DE 19901
legal@semswitch.com

https://soundos.app