User-controlled protection for the AI agent you already use.
UnoLock is not the AI model or the agent. It is the protection layer around what the user chooses to share with that agent. The goal is simple: let a user share selected data with an AI agent, protect the agent's important working state, and keep all of it aligned with UnoLock's holistic approach to user security.
Instead of trusting a plaintext API key, an exportable credential, or a random local folder full of prompts and notes, the user can give the agent a limited Space, durable encrypted memory, and a protected backup path for important local files.
You use the same UnoLock app. The difference is that you create a new Agent Key for the agent and assign it a UnoLock Space. The agent is restricted to its own protected spaces to work in, and the user keeps control of what the agent can access and do. For more information or to get started, send your agent to UnoLock AI agent setup guides.
Four Core Protections For AI Agent Work
The user stays in control while the agent gets the limited access and resilience it needs.
Encrypted Agent Memory & Secret Store
Agents get durable, user-controlled memory and a secret store through client-side encrypted cloud storage. Notes, checklists, working memory, and secrets can live in a Safe instead of plain local files.
Device-Bound Agent Access
Rather than relying on plaintext or exportable credentials, UnoLock can bind agent access to TPM, Secure Enclave, or another platform-backed non-exportable key store when the host supports it.
Separate Agent Space With Limited Keys
The agent gets its own dedicated key and its own Space, with view-only or limited edit access instead of full Safe control. Trust does not depend on giving the agent a permanent all-access secret.
Sync Backup For Local Agent Files
A one-way local-to-cloud backup protects selected agent working files such as prompts, notes, identity files, and project state. After a reset or clean reinstall, the user can manually restore what matters.
Why This Protects Both Sides
Protect The User
No broad reusable secretThe user does not need to hand the agent a permanent all-access token. UnoLock keeps the access model scoped to the granted Space and the operations allowed by that key.
Read-only agent keys stay read-only. Locked records stay locked. The Safe remains under the user's control, not the agent's convenience.
Protect The Agent
Durable encrypted working stateAgents also need protection from loss. Notes, checklists, selected files, and working context can survive host resets or reinstalls instead of disappearing with local temp files and chat history.
The result is safer continuity: important state lives in an encrypted Safe, not in a plain local directory or a forgotten terminal buffer.
Protect The Relationship
Clear limits and cleaner consentThe user can give the agent exactly the access needed for the job and explain that boundary clearly. The agent can say what it can do, what it cannot do, and what additional permission would be required.
That creates a healthier model than silent overreach or hidden dependence on one reusable secret stuffed into a config file.
Built For Real Agent Workflows
The same Safe model, adapted to practical AI assistance and controlled automation.
Local Assistants & Copilots
Give the agent a dedicated Space for durable notes, checklists, and project files without exposing the rest of the Safe. Ideal for daily assistants working in a logged-in user session.
Sensitive Automation
Keep automation inside explicit Safe and Space permissions instead of hiding secrets in env vars or a runner config. The agent can read or write only what the key permits.
Recoverable Working State
Back up selected local files into UnoLock Cloud, restore them after reinstall, and keep important agent artifacts inside the same Space as the rest of the agent's protected context.
