Your data and systems stay protected
Handing an AI agent access to your systems only works if your passwords, your network and your customers' data stay safe. Aether keeps passwords encrypted and out of the agent's reach, blocks agents from your internal network, and lets you decide what is kept about each task.
How we protect you
What your IT and security team will want to know, in plain terms.
Your passwords and keys
- Encrypted with AES-256-GCM when stored, decrypted only while a task runs, and handed to the connection that needs them, never to the agent. They never appear in the portal or in logs, and each one is tested before it is saved.
Your internal network
- Agents are blocked from reaching private and internal network addresses. They can only reach the systems you connect on purpose.
Requests coming in
- Messages sent to your agents from other systems can be signed and are checked before anything runs.
Tricks hidden in messages
- Content from customers and documents is kept separate from the agent's own instructions, and every action the agent asks for is checked before it happens.
Limits on every agent
- Each agent can only use the systems it was given, and stops after a fixed number of steps, so it cannot run on indefinitely.
What is kept about each task
- Full detail, the steps without their content, or nothing at all: you choose what your data policies allow.
Who can see what
- Each person in your portal has a role that decides which projects and agents they can see and manage.
For your security review
We'll answer your security questionnaire and walk your team through how agents reach your systems, where credentials are kept and what is recorded.
Security questions
Can the agent see our passwords?
No. Passwords and API keys are encrypted when stored, decrypted only while a task runs, and given to the connection that needs them, never to the agent itself. They never appear in the portal or in logs.
Can an agent reach our internal systems?
Only the systems you connect on purpose. Agents are blocked from making requests to private or internal network addresses, the usual way automated tools are misused to reach internal infrastructure.
Can we limit what data is kept about each task?
Yes. The record of each task can keep full detail, keep the steps without their content, or keep nothing, depending on what your data policies allow.
Do you support single sign-on?
Your portal uses secure sign-in with a role for each person. Single sign-on is on the way; ask us about timing.
Ready to put your first agent into production?
Bring us one process. We’ll map the skills, rules and integrations, and help you choose cloud, on-premises or a managed service.