Skip to main content
Your team pastes code, customer records, and credentials into AI tools every day. Shadow AI catches that content on the employee’s own device, the moment before it’s sent — and blocks, redacts, or allows it according to your policy. To decide, the agent sends the prompt to the PromptGuard engine — in our cloud, or on servers you run — and the AI tool only receives what the verdict allows.
Today the engine receives the prompt unmasked: on-device masking protects the hop to the AI tool, not the hop to the engine. If prompt text must not leave your network, run the engine yourself — see Deployment Modes. Masking on the device before the engine sees anything is on the roadmap.
It’s the same PromptGuard engine, policies, and dashboard you already use for your own apps — now watching the AI tools your employees use, too.

​
How it works

Every paste, prompt, or upload to a known AI tool gets one verdict, in milliseconds:

Block — secrets

API keys, cloud credentials, and tokens never reach the AI tool.
AWS_SECRET=… → stopped.

Redact — PII

Emails, phone numbers, SSNs, and card numbers are masked on-device, so the employee still gets help and the AI tool never receives the raw value.
jane@acme.co → [EMAIL].

Block — attacks

Prompt-injection and jailbreak payloads are caught before they’re sent.

Allow — real work

Everything else passes through untouched, with zero added friction.
Every verdict lands in your PromptGuard dashboard, tagged by surface (desktop · browser · sdk · proxy) — so application traffic and employee traffic show up in one audit trail.

​
Get started

Quickstart — protect one machine

Install, connect, and see your first block in about five minutes — on macOS, Windows, or Linux. One agent covers both AI APIs (Cursor, IDE assistants) and the AI web apps in the browser (ChatGPT, Claude) — no extension needed.

Roll it out to your fleet

Enroll many devices under one org, each with a scoped, individually revocable credential — and see the whole fleet in one place.

​
One engine, every surface

Shadow AI isn’t a separate product. It’s additional places we watch feeding the same detection engine, the same policies, and the same dashboard:

Your apps

Gateway & SDK traffic

Your team

Desktop agent + browser

One dashboard

Scan · block · redact · audit
Adding a new provider or site is a one-line change — not a new product to learn or deploy.

Run it your way

Cloud, Hybrid (your engine, our dashboard), or Air-gapped (engine and dashboard both on your servers).

​
Licensing

Shadow AI follows the same plan model as the rest of PromptGuard. There are three ways to license it:

Personal

Included with every plan. Protects one of your own devices, metered against your existing request quota — no separate bill.

Fleet

Available on the per-seat Team plans, and on those alone — a gateway plan does not grant it however large. Adds MDM enforcement, an org-wide required policy, multi-device enrollment, and a per-employee usage rollup.

Standalone

Priced per seat for rolling Shadow AI out to a whole team, or using it on its own. Cloud, hybrid self-hosted, or air-gapped.
See the pricing page for plan limits and how Shadow AI usage is metered.

​
What Shadow AI does — and doesn’t

We’d rather be upfront than over-promise:
Detection requires reading the request, so the agent terminates TLS using a certificate it installs on the device — user-approved for individuals, or MDM-managed for enterprise fleets. The agent reads the request on the device and sends its content to the engine for a verdict; the engine records the verdict with a masked preview.
We watch a curated, growing set of AI providers and web apps. New endpoints are added over time; a tool we don’t yet recognize passes through untouched.
A few apps pin their own certificate and bypass any inspection proxy. Those are out of scope by design.
The agent fails open: if the engine can’t be reached, or hasn’t answered within 30 seconds, the prompt is sent without an engine verdict. The on-device secret and PII masking still applies; prompt-injection detection does not. This keeps AI tools usable during an outage — see Troubleshooting.