Zotniq is runtime AI usage control for the endpoint. It inspects prompts to ChatGPT, Copilot, Claude, Gemini and other AI tools before they leave the employee laptop, and applies team-level rules so sensitive customer data, PII, PHI, financial records, API secrets and proprietary code don’t reach third-party AI services.
Built for CISOs, DPOs and AI governance teams at mid-market and enterprise organizations in financial services, healthcare and SaaS that are adopting AI tools without a data governance layer underneath.
Zotniq runs on the device, not the network, so policies, evidence and audit trails stay in force on and off VPN.
Core capabilities:
* Real-time detection of PII, PHI, PCI data and API secrets, plus custom data types for source code and internal identifiers
* Pre-send masking: sensitive data is replaced before the prompt reaches the AI tool
* Monitor mode to observe AI usage before enforcing, Enforce mode to mask or block
* Stop Copy Paste: one team setting blocks pastes into AI apps, AI browser tabs and AI tools running in terminals
* AI inventory of the AI apps and MCP servers in use across devices
* Kill switch for emergency, organization-wide AI blocking
* Team-scoped governance rules
* SIEM forwarding (Splunk, Datadog, webhook)
* Audit evidence export for GDPR, HIPAA, SOC 2 and DPDP
* Python SDK (pip install zotniq) with detection primitives and drop-in wrappers for OpenAI and Anthropic
Common use cases: masking PII before ChatGPT and Copilot prompts leave the laptop; stopping copy-paste of sensitive data into AI tools; audit evidence for AI data handling under GDPR, HIPAA and SOC 2; Monitor-mode rollouts that produce board-ready evidence before enforcement; SDK integration for developers building AI applications.
