How AIPSAFE365 works

AIPSAFE365 combines selected-folder observations with permission decisions for cooperating AI integrations. These are separate local services: monitoring observes activity, while a connected agent asks for authorization before acting.

Set up selected-folder monitoring

1. Install the desktop app for a supported computer. Add the folders you want to observe in Settings.
2. Open Monitoring and start the service. If you configure process names, matches are candidate processes, not verified AI identities.
3. Review folder events, process observations and any coverage errors. Operating-system permissions can limit what the app observes.
4. Stop monitoring when you do not need it. Review changed settings before restarting.

Monitoring records changes. It does not intercept credential reads, stop file changes from other applications, or prove which agent changed a file.

Permissions require a cooperating integration

1. A developer integrates the source-distributed Node SDK into the agent’s action workflow. The SDK is not a published plug-and-play integration for every AI tool.
2. In the desktop app’s integration controls, start the local permission service and enroll that agent. Give the local address and token only to the trusted integration; keep tokens out of prompts and shared logs.
3. The agent submits the action metadata, follows the permission result, and consumes authorization before acting. Review any pending request in the desktop app.
4. Allow once, block, or save an eligible exact-action exception. Review saved rules and revoke an agent’s token when its integration is no longer needed.

Installing AIPSAFE365 or adding a process name does not connect an agent. Unintegrated applications can bypass this workflow. The Ollama text-file adapter is a source-distributed developer preview for a limited new-text-file action, not a bundled customer integration.

Optional Pro checks use information you submit

With Pro monitoring enabled, cloud checks evaluate up to 10 saved domains roughly hourly, including while your computer is off. The dashboard shows check history and summaries. You can also choose to submit text for scam-pattern review.

Checks use warning patterns and an operator-maintained indicator list. They do not visit website content, inspect your inbox or calls, monitor other apps, or guarantee that every threat will be detected. A result without a warning is not proof that a site is safe.

Read the published privacy notice before submitting domains or text. Disabling monitoring does not cancel a subscription.

Check your computer before downloading

• Windows: Windows 10 or 11 on 64-bit Intel or AMD PCs. The installer is unsigned and may show an unknown-publisher warning. If security settings block installation, contact support; do not turn off protection.
• Mac: Apple Silicon, M1 or newer.
• Not included: Windows ARM, Intel Mac, Android or iPhone releases. iPhone push alerts and AI Agent Defense are not available yet.

Supported desktop downloads and setup steps · FAQ and billing help

Understand registration before choosing a plan

Basic pays for account-linked registration of one computer and subscription management. Local monitoring, saved history and rules, and cooperating-agent permissions work independently of billing. Paying does not unlock additional local enforcement. Pro adds the optional cloud checks described above.

Compare Basic and Pro and read the subscription terms. For product or account questions, the published support address is teamsafe@aipsafe365.com.