Security built around explicit context.
AskAny.ai is designed so AI work starts with a user action: select text, attach a file, crop a screenshot, speak a prompt, or type a request. The trust model is simple: make context intentional, routing visible, and enterprise controls available when teams need them.
What teams can verify first
AskAny.ai processes context the user actively submits.
AskAny.ai does not train its own foundation models on customer prompts, files, screenshots, or chats.
Eligible teams can route approved model traffic through customer-controlled provider keys.
DPA/AVV, subprocessors, model-provider notes, and data-flow documentation.
User-triggered
AI requests start only when someone submits selected content or a prompt.
Visible routes
Users and admins can reason about which model/provider route is being used.
Admin controls
Enterprise admins can manage model visibility, BYOK routes, and team membership.
Reviewable assets
Trust pages support security review before broader deployment.
From selected context to answer, step by step.
The security boundary is easiest to understand when you can see where data enters, where it is prepared, and which route processes it.
User chooses context
Text, files, cropped screenshot, voice, or typed prompt.
AskAny.ai prepares request
Context, model choice, task instructions, and required metadata.
Route is checked
Enterprise workspaces can enforce approved models and BYOK routes.
Model inference
The selected provider or customer-key route processes the request.
Answer returns
Users can copy, replace, save, continue, or discard the output.
A practical security posture for desktop AI.
AskAny.ai focuses on the controls teams ask about before deployment: context capture, routing, provider behavior, business agreements, and admin enforcement.
Explore Enterprise controlsContext controls
- •User-triggered capture, not continuous recording.
- •Region-based screenshot capture is visible to the user.
- •Users submit only the text, files, voice, or prompts they choose.
Provider controls
- •Official model APIs where supported.
- •No-training options configured where available.
- •Provider matrix documents handling differences.
Enterprise controls
- •BYOK routing and model definitions for eligible workspaces.
- •Workspace-wide model/provider visibility controls.
- •Admin and member role management.
Business controls
- •DPA / AVV support for eligible customers.
- •Subprocessor list and trust documentation.
- •Custom deployment and retention reviews on request.
Security FAQ: the hard questions
Do you have SOC 2 or ISO 27001 certification?+
Not yet. AskAny.ai currently provides a security overview, DPA/AVV support, subprocessor list, and architecture/data-flow documentation. Formal certification can be prioritized based on enterprise demand.
Can my company force all traffic through our own keys?+
BYOK is available on request for Enterprise and power users. Supported deployments can route approved model traffic through your own provider account so your provider contract, audit logs, retention controls, and billing apply.
Can we restrict which models employees use?+
Yes. Using Workspace Enforcement in the Enterprise dashboard, admins can toggle off a provider or individual model so it disappears from the employee app experience.
What happens when a user submits a confidential document?+
The selected document content is processed to generate the requested answer and may be sent to the selected model provider. Users should submit only data they are authorized to process. Enterprise customers can request approved routes, retention controls, and DPA terms.
Do third-party AI providers retain prompts?+
Provider handling varies by provider, plan, and contract. AskAny.ai maintains a provider matrix and configures no-training options where supported. BYOK customers should review and configure their own provider accounts.
Can AskAny.ai be self-hosted?+
Self-hosting is not generally available as a standard plan. Custom enterprise deployment reviews may be considered for qualified customers.