Our approach
Kueso brings restaurant operations, team knowledge and AI assistance into one place. That creates a responsibility to handle business information thoughtfully, explain clearly how the service works, and make security part of the product rather than an afterthought.
Kueso is built on established infrastructure rather than a proprietary database or identity system created from scratch. Security follows a shared-responsibility model: our providers operate the underlying infrastructure under their own security programs, and Kueso is responsible for the application logic that decides what each person can do with the information stored there.
This page separates four kinds of statements, so it is always clear what each one rests on.
- In place
- Kueso application controls supported by our internal application security review.
- Provider
- Protections and certifications documented by the providers we use. These belong to the provider, not to Kueso.
- In progress
- Security improvements we are actively working on.
- Planned
- Future objectives. These are not represented as controls already in place.
Security foundations
The controls below are application-level safeguards supported by our internal application security review. They describe how Kueso is designed to behave in ordinary product workflows. We do not claim that every production database policy, storage path or alternate access route has been independently tested; validating these controls is part of our ongoing program.
Authentication and identity verification
In place
Authenticated requests to Kueso’s services use verified sign-in tokens. The service determines who is making a request from that verified token, rather than trusting a user identifier supplied by a device. Requests without valid authentication are rejected on authenticated routes.
Sign-in and identity services run on Supabase’s managed authentication infrastructure.
Application-level access controls
In place
Kueso organizes restaurant work into Spaces and channels. Application-level authorization checks are designed to limit channel messages and content to the owner and participating members. Access is determined by membership, not by knowing a channel’s identifier.
Different operational tools use permission models suited to their purpose, and documents in the application follow the applicable workspace and tool permissions.
AI permissions and human oversight
In place
Kueso’s assistant is intended to work within the permissions of the signed-in person, not as an unrestricted administrator of a restaurant’s information. For supported workflows, it retrieves information through permission-aware tools that use the signed-in user’s identity, so answers are grounded in information that person can access through the application.
Sensitive actions, such as sending messages or emails and making certain operational changes, use confirmation flows. Kueso presents a draft for review before the action is carried out. Routine read-only assistance does not require the same confirmation.
Employee information safeguards
In place
Employee Intelligence is designed around independently owned staff directories. It keeps staff profiles separate from app accounts, uses owner-managed access with explicit manager invitations, and reserves compensation information for the directory owner. When a roster is imported, managers review the extracted information before it is saved.
These are product controls of the Employee Intelligence module and apply where that module is enabled. Availability of individual tools may vary by deployment.
Connected account protections
In place
Integrations such as Slack and Gmail are optional. They are connected by users to enable specific workflows, and they are relevant only when a customer chooses to connect them. Supported outbound messages sent through a connected account require confirmation before they are sent.
Operational records and temperature logs
In place
Temperature Logs is designed to preserve submitted readings and to keep versions of published log templates, which helps retain the history and context of each operational check. Participation is managed by the log’s creator, with controlled permissions for who can submit readings.
Kueso also helps teams organize restaurant documents, procedures and operational knowledge so authorized users can find answers without searching disconnected systems. Document access follows workspace and tool permissions. The privacy of shared file links is a separate area of ongoing verification and improvement.
These records support, but do not replace, a restaurant’s own food-safety procedures, training and regulatory responsibilities.
Payments and payment providers
In place
Kueso uses Stripe and Apple for supported payment flows. Kueso does not collect payment card numbers directly in its application.
AI processing and data handling
Provider
Kueso uses Anthropic and OpenAI for AI capabilities. To process a request, relevant prompts, conversation context and retrieved material may be sent to the provider needed for that request. This is an important part of how the service works, and we believe customers should understand it.
Anthropic maintains a public Trust Center with compliance documentation and controls for its API services. OpenAI states that data sent to its API is not used to train its models by default unless a customer explicitly opts in; separate abuse-monitoring and feature-specific retention terms may apply.
These statements describe the providers’ published policies. They are not an independently verified description of a Kueso-specific retention configuration.