Keep Company Data Safe While Using AI.

AI DLP stops sensitive data - PII, credentials, source code, customer records, internal documents - from leaving your network when employees paste into ChatGPT, Claude, Copilot, Gemini, or any AI tool. NexusNest is different because we don't ship a browser plugin or rely on the AI provider's API. We install a tiny agent on the employee's machine that intercepts AI traffic, redacts sensitive fields before the request reaches the AI provider, and forwards a redacted version to the AI. The AI never sees your secrets.

PromptWall is the redaction engine - it intercepts AI requests and redacts sensitive data. NetLens is the visibility layer - it shows all employee AI activity, who's using which tools, and which prompts triggered redaction. They ship as a single agent. PromptWall blocks leaks; NetLens proves they were blocked and gives you the audit trail.

ChatGPT (chat.openai.com, chatgpt.com, api.openai.com), Anthropic Claude (claude.ai, api.anthropic.com), Google Gemini (gemini.google.com, generativelanguage.googleapis.com), GitHub Copilot, Cursor, Codeium, and any other AI tool you add to the allowed list. The agent inspects HTTPS traffic to known AI endpoints; you can add custom domains as your team adopts new tools.

Yes. The agent works at the OS network layer, not in the browser. It catches prompts from desktop AI apps (ChatGPT, Claude, Cursor, JetBrains AI), browser extensions (Copilot, Continue), CLI tools (Claude Code, OpenAI CLI), and SaaS chat tabs. If it goes out as HTTPS to a known AI domain, we see it and redact it.

Personal information (names, emails, phone numbers, addresses, government IDs), financial data (credit cards, bank accounts, IBANs), health data (medical record numbers, diagnoses), credentials (API keys, OAuth tokens, AWS access keys, JWTs, passwords), source code blocks, and internal document markers. You can also define custom regex or keyword policies for industry-specific data (PCN, MRN, MR-N, claim IDs, etc.).

No. The agent adds under 5 milliseconds of latency per prompt - imperceptible compared to AI response time which is usually 500ms+. Redaction happens in parallel with the prompt being forwarded. If the redaction server is unreachable, the agent fails open: prompts go through unredacted and a non-blocking degraded-mode banner appears, so employees never lose access to their tools.

Sign up at app.nexusnest.io, add employees, generate a license key for each, and share the installer link. Employees run the .pkg (macOS), .exe (Windows), or .deb (Linux) installer once, paste their key, and they're protected. Average install time is under two minutes. MDM-friendly silent install flags are available for Jamf, Intune, and Kandji.

Employees see a one-time menu-bar banner when sensitive data is redacted, telling them the action was successful. They cannot disable redaction from their device - the agent runs as a launchd / systemd service and the proxy is system-wide. Admins can disable redaction globally from the dashboard if needed (e.g., during onboarding).

No. We never store the original prompt - it's discarded the moment the redacted version is computed, so the AI provider only ever receives the redacted text. Only the redacted prompt (the version the AI provider received), redaction metadata (which categories triggered), and timestamps are sent to the NexusNest dashboard. Request and response bodies are NOT stored by default; you can opt in per-domain for specific compliance use cases.

The standard SaaS deployment fits most companies. For air-gapped, regulated, or fully on-prem requirements, contact us at hello@nexusnest.io - we can ship a self-hosted server image (Docker / Kubernetes) that runs entirely inside your infrastructure with no outbound dependencies.

DPDP, GDPR, HIPAA, and similar regulations require demonstrable controls over how personal data leaves the organization. NexusNest gives you the technical control (redaction before egress) plus the audit log (which categories were redacted, when, for which employee). We don't issue compliance certifications, but we provide the evidentiary trail your auditors will ask for.

Starter is ₹499/seat/month with 500 redacted prompts per seat. Business is ₹999/seat/month with 2,000 prompts/seat plus custom policies, payload storage, SSO, and Slack alerts. Enterprise is ₹1,999/seat/month with 6,000 prompts, priority support, and an SLA. Yearly billing is 20% off. Every plan includes the desktop agent and the dashboard.

Yes. Every new account gets a 14-day trial with all Enterprise features unlocked and a 5,000-prompt quota - no credit card required. After the trial you pick a plan from the dashboard.