See it. Detect it. Protect against it.
Openwake shows which AI your people and machines are using, flags the risky use, and can stop a secret before it is sent.
See which AI is in use, and where.
Add a machine from your Openwake account and it starts reporting within minutes. It runs on Windows, macOS and Linux, and you can roll it out by hand or with the tools you already use, like Intune or Jamf.
- Which AI tools are installed, on which machines, for which people
- Claude Code sessions, and what they connect to
- The MCP servers each tool can reach
- Which AI providers each machine really talks to
Know when something risky happens.
A password pasted into a chatbot. Customer data going to a personal account. An AI tool nobody approved. Openwake raises each one as a detection that tells you who, on which machine, in which AI, and what kind of data was involved.
- Set policy by AI tool, by personal or company account, and by kind of data
- Try a policy against past activity before you turn it on
- Alerts by email and Slack
- Keeps labels, not what people wrote, unless you turn on the Evidence Vault
Stop a secret before it is sent.
In supported Chromium browsers such as Chrome and Edge, Openwake checks what is about to go to an AI site. If it is a password or an API key, it can warn the person or block it, and tell them why.
- You choose: observe, warn or block
- It knows a personal account from a company one
- Every block, and every warning the person chose to override, is recorded
On Windows, the browser sensor arrives with the endpoint sensor, managed or not. On a Mac with the endpoint sensor, add Scout for Browser from the Chrome Web Store once and it enrolls itself. Firefox and Safari are not supported.
Try a policy exampleTry a browser policy.
Send a sample credential, then explore the policy modes.
Live simulation in your browser, using the same patterns and policy shape the real sensor uses. Only a credential is ever blocked; company data and source code are a warning, never a block. AKIAIOSFODNN7EXAMPLE is AWS's documentation example, not a real key.
Declared. Observed.
See what the provider publicly promises and what is actually happening on your endpoints, in one record. The gap between the two is what a policy is for.
Declared
What providers say they do, in their own documentation.
- Data usage
- Training
- Subprocessors
- Security claims
Observed
What actually happens, in your environment.
- Real data flows
- Actual service usage
- Downstream destinations
- Policy violations
Sensitive data can travel further than you expect.
ChatGPT Enterprise and personal ChatGPT look almost identical from the network. The risk changes with the account, the data being sent, and what that provider does with it. Openwake connects the person, the account, the data class and the provider's own terms, and the same record comes back wherever AI gets used: the person, the machine, the application or agent, the tools it called, the provider, and everyone that provider sends the data on to.
- PersonPerson
- MachineDevice
- Application or agentApplication
Browser, CLI, runtime - ToolsTool / MCP
MCP servers - ProviderProvider
- ProcessorsKnown processors
Labels, not content.
No keystrokes. No screenshots. Openwake keeps what kind of data was involved, not what people wrote, unless an owner turns on the Evidence Vault. Read the privacy policy.