iDox.ai stops sensitive data from leaking through AI tools and keeps AI agents inside your rules — so your people can adopt AI faster, with total trust.
The key to information security in an autonomous AI world.
Every paste, upload and automated action is a chance for confidential data to leave your control. Once it’s out, you can’t take it back.
of AI users paste data into AI tools — and 22% of those pastes contain PII or payment data.
Source: LayerXof the data employees paste into ChatGPT is confidential company data.
Source: Cyberhavenof all AI prompts contain sensitive data, often without anyone realizing.
Source: HarmonicAfter employees pasted sensitive internal code into ChatGPT, Samsung restricted generative AI tools on company devices and internal networks.
Why it matters: AI agents now buy, send and install things on their own — the risk is shifting from what AI sees to what AI does.
Source: TechCrunch · May 2, 2023 ↗Every employee and every AI agent gets a governed path to use AI — instead of a blanket ban or a leap of faith.
Sensitive details are found and replaced in real time — on the device, before anything leaves. AI still gets the context it needs to be useful, and never the data it shouldn’t.
Agents can work at full speed. When one tries something risky — a payment, an install, sending files outside — it is examined first, then allowed, sanitized, held for approval or blocked.
Every interaction leaves a who-what-when record. Documents leave with an audit certificate. Backed by SOC 2 and ISO 27001, and built for GDPR, HIPAA, CPRA, FOIA and FISMA programs.
Data is what AI receives. Intent is what AI may do. iDox.ai protects both — in documents, in prompts and in the actions AI agents take.
Every action an AI agent takes stays within authorized purposes, permissions and company policy.
“Renew our software subscription.”
Purchase request caught
$25,000 is over the $10,000 limit
Approval required
Only the approved plan renews
10 high-risk intents examined — in text, code, documents and system commands
Switch between a document and an AI prompt to see how the same sensitive data is protected.
Patient Jane Doe (DOB 04/12/1986) was admitted to St. Mary’s Hospital. Contact: [email protected] / (415) 555-0132. Insurance no. HX-4471-9082.
iDox.ai highlights every sensitive item it detects.
Safe AI, tailored to the regulations and workflows of your industry.
Privileged and personal information is redacted in bulk, with a certificate for the production record — and case facts never reach a public AI tool while associates draft.
Weeks of manual page-by-page review, inconsistent redactions and a risk of privileged data in AI prompts.
Consistent, auditable redaction in hours — and a policy that follows every associate’s AI use.
Patient identifiers are protected while the documents stay readable for analysis, supporting HIPAA and GDPR obligations — and PHI never ends up in an AI assistant.
Identifiers hand-removed from every record; PHI at risk of ending up in an AI chat.
Documents anonymized at scale and shared with confidence; PHI sanitized on-device.
Redactions are consistent, exemptions are labeled and outputs are audit-ready — and any AI use stays inside agency policy.
Backlogs, manual redaction and uneven decisions across reviewers.
Faster, repeatable responses with a complete record of what was redacted and why.
Account and personal data are protected before the AI sees them, card numbers can’t be pasted, and an agent’s purchase above a user’s limit waits for approval.
Choose between AI’s productivity and customer-data risk — or ban AI altogether.
Teams use AI with guardrails: data protected, spending bounded, everything logged.
Tenant personal and financial details are removed from contracts and transaction records, and stay protected when staff use AI to summarize or draft.
Sensitive tenant data scattered across contracts and email threads.
Clean, shareable documents and a safe path to use AI across the property lifecycle.
Our AI redaction software automatically identifies and redacts sensitive data with industry-leading accuracy to help ensure nothing is missed.
Reduce document review and redaction time from hours to minutes using automated redaction tools and intelligent workflows.
Minimize costs associated with manual document review, detection, and redaction through scalable redaction software.
Give legal, government, healthcare, and compliance teams valuable hours back with AI-powered redaction tools that automate repetitive tasks.
“Quick, well designed, and saved our team over 97% of time. Our team was able to meet subject access requests and meet GDPR mandates using the iDox.ai Suite solution.”
Carl BevanRoyal College of Nursing“iDox.ai Suite saves the H Law Group 10 to 15 minutes per document which amounts to over 350 hours annual time savings!”
H Law GroupLegalBacked by SOC 2 and ISO 27001 certifications, built to help you meet GDPR, HIPAA, CPRA, FOIA and FISMA requirements.
Independently certified security controls.
AES-256 at rest, SSL 2048-bit in transit.
iDox.ai checks data locally — it never has to leave your machine.
Hosted service or your own environment, with SSO and audit logs.
One platform, three ways in. Start with the one that fits your first use case.
Redact, anonymize and compare documents at scale, with an audit-ready certificate.
Explore Suite →
Real-time, on-device protection for what people and AI agents send to and do with AI.
Explore Guardrail →
Set policy once, deploy everywhere and see every event across your organization.
Explore Enterprise →Transparent plans for individuals and teams, and custom editions for the enterprise.
Real-time, on-device protection for pros and small teams on Windows, macOS and Android.
Try Guardrail freeStarter $39/mo (5,000 pages/yr) · Premium $89/mo (50,000 pages/yr, co-review). 7-day free trial.
Management Console, SSO, industry editions, cloud or on-premises deployment. Minimum 3 users.
iDox.ai makes enterprise AI safe, private and governable. It protects sensitive information in documents and while people and AI agents use AI — masking or blocking data before it leaves, holding risky agent actions for approval, and recording every decision for audit.
Give them a governed path instead of a ban. iDox.ai checks prompts, uploads and files on the device before they reach an AI tool, masks or blocks sensitive data such as personal information, financial data, credentials and source code, and lets safe content continue — so AI stays useful.
Governing AI intent means understanding what a user or AI agent is trying to accomplish and making sure its actions stay within authorized purposes, permissions and company policy. Instead of only detecting a risky instruction, it controls how that instruction becomes an action: intercept, examine, enforce, then monitor and verify.
Traditional DLP looks for known patterns after content is already moving. iDox.ai protects data at the moment of interaction with AI and understands meaning and intent, so it can catch sensitive data and unsafe instructions that match no signature.
iDox.ai works at the point of use, so it is designed to cover chat assistants, productivity copilots, coding agents, custom and open-source models, and AI agent tool calls, rather than being tied to one vendor.
No. Detection and sanitization happen on your device in real time, so sensitive data does not have to leave your machine for iDox.ai to protect it.
iDox.ai supports GDPR, HIPAA, CPRA/CCPA, FOIA and FISMA requirements, and is SOC 2 and ISO 27001 certified. Technology controls complement, but do not replace, legal and organizational measures.
iDox.ai Suite starts at $39 per month (Starter) and $89 per month (Premium); Guardrail Individual is free for a limited time; Enterprise and industry editions are priced by contact. See the pricing page for details.
See how iDox.ai lets your whole organization use AI — safely, responsibly and confidently.
