iDox.ai Guardrail is an on-device AI security app that blocks sensitive data from reaching ChatGPT and stops unsafe AI agent actions in real time.
About
iDox.ai Guardrail is a security application from iDox.ai that prevents sensitive data from leaking into AI tools such as ChatGPT. The problem it addresses is well documented: employees routinely paste confidential information into AI chatbots, and once submitted, that data cannot be retrieved. Guardrail aims to stop those leaks before they happen rather than detect them afterward.
The app runs quietly in the background on a user's device and inspects everything being sent to AI services. It operates on three layers: it masks or blocks sensitive text and files before they reach an AI chat, it detects prompt injection and unsafe instructions, and it monitors the communication between AI agents to block unauthorized commands. iDox.ai, the company behind the product, is a B2B SaaS provider focused on data privacy and AI safety, serving government, legal, financial services, life science, and real estate customers from its headquarters in Fremont, California.
Key Features
Pricing & Fees
- Pricing: The Individual edition is listed as free for a limited time with 7-day full access and no credit card required; Enterprise pricing is not publicly listed on the pages we reviewed; check the official site for current plans.
- Editions: Two tiers are offered — Individual (on-device protection for professionals and small teams) and Enterprise (centralized governance with policy packs for GDPR, HIPAA, FOIA, and other regulations).
- Platform availability: Windows 10 and 11 (64-bit), macOS on Apple Silicon (M1 and later), and Android via Google Play; Windows power users can also install with a single PowerShell command.
- On-device processing: Interception runs locally on the device rather than sending data to the cloud for inspection.
- Compliance alignment: The product page highlights GDPR, HIPAA, FOIA, SOC 2, and ISO 27001 among its compliance frameworks.
Who It's For
iDox.ai Guardrail is built for two audiences. Individuals and small teams who use ChatGPT or AI agents in daily work can install the Individual edition to keep client details, source code, and personal data out of AI services. Enterprise IT and security teams, along with compliance officers in regulated industries such as legal, finance, healthcare, and government, can deploy the Enterprise edition to enforce central AI usage policy, audit agent activity, and meet obligations under GDPR, HIPAA, FOIA, and similar regimes.
FAQ
What is iDox.ai Guardrail?It is an on-device application that intercepts sensitive data before it reaches AI tools like ChatGPT, detects prompt injection, and monitors AI agents for unauthorized commands.
Is Guardrail free?The Individual edition is listed as free for a limited time with 7-day full access and no credit card required; Enterprise plans require contacting the vendor.
What platforms does it run on?It supports Windows 10 and 11 (64-bit), macOS on Apple Silicon Macs (M1 and later), and Android, with real-time alerts on mobile.
Does my data get sent to the cloud to be checked?No — the product's core interception runs on the device, so content is inspected locally rather than uploaded for analysis.
What is the difference between the Individual and Enterprise editions?The Individual edition provides on-device protection and agent auditing for one user or small team, while the Enterprise edition adds a management console, role-based access, a full audit trail, and cloud or on-premises deployment.
Related Tutorials
View all →
Related Tools
View all →
viAct Job Hazard Analysis (JHA) Software AI-powered Job Hazard Analysis and Job Safety Analysis software from viAct that automates risk assessments, standardizes JHA workflows and supports EHS compliance. 
Quivr Quivr is a free, open-source news search and monitoring engine that ingests news feeds, mailboxes and social posts, keeps every version, and alerts teams when a followed topic appears. 
Yuxi - RAG Harness for AI Agents Yuxi provides a RAG harness to enhance AI agents with retrieval-augmented generation capabilities. 
omp - A Coding Agent with the IDE Wired In Oh My Pi (omp) is a coding agent that integrates with your IDE, leveraging LSP and DAP for seamless development. Built in Rust for performance. 
DeepEval by Confident AI - LLM Evaluation Framework Open-source LLM evaluation framework with 50+ plug-and-play metrics for AI agents, RAG, chatbots, and more.