Containment AI

Protect your organization from data leaks to AI chatbots

Containment.AI governs how your team uses AI — in the browser, in real time. Containment.AI is the connected tier of a single governance discipline: the Mission Authorization Gateway stands between AI intent and mission effect, and the browser extension extends that same enforcement to everyday employee AI use. It inspects prompts as your team types them into AI chat assistants, including ChatGPT and Claude, and blocks policy violations before they reach the AI provider. What it does: AI Chat Firewall — inspect employee AI prompts in real time and block policy violations before submission. Deterministic enforcement by default — allow/block decisions come from your organization's rule-based policies, not a model guess. An optional AI-assisted prompt-injection check stays off unless your administrator turns it on. Audit trail — every policy violation is recorded with metadata (policy, severity, matched pattern, site, and time) designed to support GDPR and EU AI Act Article 12 record-keeping obligations. Central policy control — administrators configure policies once in the Containment.AI dashboard; the extension enforces them across every connected AI surface. Fail-closed by default — when the extension can't reach policy, it blocks rather than leaks. Built for regulated and security-conscious teams — defense, aerospace, critical infrastructure, and any organization that needs a defensible record of how its people use AI. Install, sign in to your organization, and your policies are enforced in the browser immediately. For IT and security teams: deploy at scale via Chrome Enterprise group policy or MDM. Pairs with the Containment.AI governance dashboard for audit logs, policy management, and compliance reporting. How it works: your IT admin configures policies in the Containment.AI dashboard. This extension pulls your organization's policy configuration to your device. When you submit a prompt on a monitored AI site, the extension sends the prompt text over an encrypted connection to Containment.AI's policy-check service, which evaluates it against your organization's policies and returns an allow/block decision before the prompt reaches the AI provider. Privacy-first design: - Prompt content is transmitted only to Containment.AI's policy-check service, over TLS, solely to evaluate it against your organization's policies — never to advertisers or data brokers. - Full prompt content is not stored by the policy check; only violation alert metadata (policy name, severity, violation message, and the matched term or pattern) is retained for your organization's audit log. - Sessions stored in chrome.storage.local (per-device, not synced). - Source code available under proprietary-evaluation license. Requires a Containment.AI account (free tier available at https://app.containment.ai).