BrowzWork Security Model
Last reviewed: October 9, 2026
BrowzWork provides an execution and access-control layer between supported AI clients and user-controlled devices. Security controls address authentication, authorization and execution boundaries; they do not verify that every decision made by an AI client is correct.
Device ownership and pairing
Devices are installed and paired by their owners or authorized operators. BrowzWork uses authenticated device connections and account association checks to route supported requests to the appropriate Local Agent. Pairing is not a guarantee that every future AI-generated operation will be harmless.
Workspace access
Workspaces are managed locally by the Agent and access is subject to supported authorization controls. For personal accounts, added Workspaces may be available to connected AI clients operating under the Device Owner's authorized account. Workspace access is not a general-purpose operating-system sandbox.
Commands, scripts and their child processes can have indirect effects consistent with the operating-system permissions and security restrictions actually enforced on the device.
OS Access
OS Access is separate from ordinary Workspace access and requires explicit local enablement where supported. It may allow operations beyond added Workspaces. Enable it only when broader device access is necessary and understood.
AI clients and automatic execution
AI clients select or generate requests. BrowzWork does not guarantee their instructions are accurate or appropriate. Within enabled access scopes, supported operations may run without individual confirmation. Existing restrictions are designed to reduce risk, but not every incorrect or destructive side effect can be prevented. Operations may be irreversible.
Keep independent, tested backups and take particular care with sensitive data and production environments. BrowzWork is not a backup or guaranteed recovery service.
Account and connection security
Protect both your BrowzWork account and the third-party AI accounts you connect. Use multi-factor authentication where supported, review paired devices and local access settings, and disconnect clients you no longer trust.
Report a security issue
Report suspected security vulnerabilities privately to support@browz.work with the affected component and potential impact. Do not access other users' information or disrupt production systems while investigating.
Scope of security assurances
BrowzWork is designed to apply supported authentication, authorization and execution controls. Those controls do not guarantee error-free AI behavior, fully reversible commands, uninterrupted services or prevention of every data loss. See the Terms of Service and Privacy Policy for the applicable contractual and data-handling descriptions.