The platform
See the AI at work.
Know who answers for it.
YuviRed is an AI visibility and governance platform that you run on your own Windows server. It looks for AI tools and agents on your Windows computers, checks what people send to AI sites, and gives people the context and the controls to decide what happens next.
01 · AI visibility
Which AI is in use, and where?
The AI Map draws every connection YuviRed has found: the computer, the application on it, and the AI provider it reaches. Select any one of them to isolate its path and read its detail.
Looks for AI websites in the browser, AI browser extensions, AI tools on the computer and command-line coding assistants
An agent on each Windows computer, added from the console with Add Computer
Each finding tied to a computer and, where known, a user
02 · AI agents (AMS)
Which AI agents are running, and who answers for them?
AMS, the Agent Monitoring System, finds the AI agents running on your computers and gives each one a workspace: where it runs, who runs it, what it did, and what you can do about it right now. The record keeps what your team declares, such as the owner, the manager and the business purpose, beside what the computer reports.
Owner, manager, department, purpose and approval on record
Hold, Stop now and Pause, shown as done only when the computer confirms
Rules per runtime, computer, user or agent, starting in Monitor only
Coverage gaps shown, never hidden
03 · Investigation
What happened, and what was done about it?
Each alert shows who, what, where, when and why, with a severity, a status, an owner and the number of days it was seen. Analysts can add notes, record a ticket reference, and see when a finding they closed comes back.
Evidence and a plain-language reason in every finding
Assignment, status and analyst notes
A flag when a closed finding returns
04 · Prompt Protection
What are people sending to AI sites?
In Chrome and Edge, the YuviRed extension checks prompts and attachments on supported AI sites, including ChatGPT, Claude, Gemini, Microsoft Copilot (copilot.microsoft.com) and Perplexity, before they are sent, and applies your policy. Prompt Monitoring records what was found and what happened, with a masked sample instead of the text.
Seven presets, from Monitor Only to Maximum, or your own policy
Warnings ask the person to decide; blocks hold the prompt
No certificates installed, no TLS interception, no keystroke logging
Supported controls
Decisions stay with people. The product carries them out.
YuviRed acts only on the rules your administrators set. A new installation starts in Monitor Only for Prompt Protection and in Monitor only for AMS, so nothing is warned about, blocked or stopped automatically until you choose. AMS interventions and website restrictions ask for a reason, and every administrative change is recorded in the audit log.
01
Classification
Every AI tool YuviRed finds is given a classification from a catalog you can extend, and your team can change it. The classification decides how findings are labeled and how urgently they are raised. It does not change anything on the computer.
02
Review and approval
Each agent in the register can be taken through review and approval, with the decision, the person who made it and the date on record.
03
Prompt Protection
Choose a preset or build your own policy: for each kind of sensitive information, record it, warn and ask, or block it. Every saved policy is a numbered revision you can test first and restore later.
04
AI agent controls
Hold, Stop now, Pause and Resume work through the YuviRed collector on the computer and read as requested until the computer confirms. With a YuviRed connector turned on, an administrator can also cancel a task and, under Enforce, approve or deny one tool call an Ask rule holds; messages reach Claude Code and agents built with the YuviRed SDK. Each control says whether it is available right now, and why not.
05
Website restrictions
On the Windows computers you choose, YuviRed can add an AI website to the blocked-sites policy in Chrome and Edge. One person requests the change and a second administrator approves it; where only one administrator exists, that administrator confirms it with a recorded reason. Every step is recorded, and the change can be removed again.
06
Audit record
The audit log, the record of who did what in YuviRed, can be exported as CSV, JSON or Common Event Format (CEF) through the YuviRed API, for the log tools you already use.
How it runs
On your server, in your environment.
A server you operate
YuviRed Lite installs on a 64-bit Windows computer you operate. A dedicated virtual machine with 4 CPU cores, 8 GB of memory and 64 GB of disk is recommended.
An agent on each Windows computer
Add Computer makes an installer for the Windows computers you choose. Each computer’s agent looks for AI tools and agents and reports to your server.
Stored where you run it
What YuviRed collects from your computers is stored on your server.
Watch it in the product.
Three short narrated walkthroughs, recorded from the product on synthetic sample data.
See it on your own computers.
Download YuviRed Lite free, or compare the editions to find the capacity you need.















