Managed cloud
Use leezy’s managed service with primary application storage in the EU and EU LLM routing to eligible EU model deployments. Review integrations and other subprocessors separately for your use case.
AI agents with EU model processing
Automate customer conversations with primary application storage and configured AI model processing in the EU. See how the managed service routes requests, which providers are involved, and where your responsibilities begin.
- EU
- Primary application data storage
- EU models
- Managed requests to eligible EU deployments
- Encrypted
- In transit and at rest
- Private
- Dedicated and customer-managed options
Data flow
Know where data goes, before you deploy
A leezy agent retrieves approved context, prepares the relevant prompt, sends it to the configured model, and returns the answer to the user.
- 01
Approved sources
The agent retrieves only the knowledge and connected-system context needed for the request.
- 02
Application storage
Knowledge, configuration, and conversation records are stored in the primary application environment in the EU.
- 03
Model inference
The prompt goes through an EU LLM gateway to an eligible EU model deployment. Embeddings for knowledge search use Azure France Central. Private deployments can have their own agreed model path.
- 04
Grounded answer
The response returns through leezy, with the conversation and operational record handled according to the selected setup.
Managed model requests use an EU LLM gateway. Eligible models are restricted by EU regional deployments or EU location metadata in the model catalogue, alongside leezy’s reviewed model list. The gateway alone does not determine inference location. Integrations and other service providers have separate processing arrangements.
Deployment options
Match the operating model to your risk profile
Start with managed infrastructure or define a more isolated setup with our enterprise team.
Dedicated environment
Agree an isolated single-tenant setup with dedicated capacity, storage, domain, and retention requirements.
Your cloud or on-premise
Deploy in your AWS, Azure, or GCP account, or on your own servers, with model routing and integrations designed around your controls.
Security controls
Practical safeguards for day-to-day operations
Technical controls support your security process, while configuration and organizational responsibilities remain explicit.
Encryption
Data is encrypted in transit with TLS and at rest through the infrastructure providers used by the selected deployment.
Role-based access
Define who can manage agents, knowledge, conversations, and workspace settings within your organization.
Auditable changes
Track relevant actions and configuration changes so teams can review who changed what and when.
Retention controls
Set retention and deletion requirements according to the product plan and agreed deployment configuration.
Responsibilities & safeguards
A clear role for every party
Data protection depends on the full processing context. These are the standard roles and safeguards described in leezy's privacy information.
- Customer as controller
- Your organization decides why and how personal data is processed through the agent and configures the use case accordingly.
- hypescale as processor
- hypescale GmbH operates leezy and processes customer data under the agreed data processing terms.
- Documented subprocessors
- Infrastructure and model providers are documented so your legal and security teams can review the processing chain.
- Transfer safeguards
- Where a provider processes data outside the EEA or the United Kingdom, contractual safeguards such as Standard Contractual Clauses apply where required.
Questions security teams usually ask
Is leezy GDPR-compliant?
GDPR compliance depends on your purpose, legal basis, configuration and organizational measures. leezy’s managed service combines primary EU storage with configured EU model processing. These locations support your assessment; they do not replace a review of processing terms, subprocessors, retention and your responsibilities as controller.
Where is customer data stored?
Primary application data is stored in the EU. In the managed service, model requests go through an EU LLM gateway to eligible EU deployments; knowledge-search embeddings use Azure France Central. Storage, model processing and connected integrations are separate parts of the data flow.
Is our content used to train public AI models?
leezy’s privacy policy states that customer content is not used to train third-party AI models. Training use is separate from provider retention for abuse monitoring. The applicable provider terms determine retention for the configured service; EU routing alone does not establish zero retention.
Can data be processed outside Europe?
The managed model path is configured for EU processing through EU LLM routing and eligible EU model deployments. Connected integrations and other subprocessors may have processing outside the EEA or the United Kingdom. Where required, transfer safeguards such as Standard Contractual Clauses apply. EU model routing does not establish whole-service EU residency.
How is data encrypted?
Data is encrypted in transit using TLS and at rest through the cloud infrastructure used by the selected deployment. The exact controls can differ between managed, dedicated, and customer-operated setups.
Can leezy run in our own environment?
Yes. Enterprise deployments can be designed for your AWS, Azure, or GCP account, or for on-premise infrastructure. Model routing, integrations, operations, and support are defined with your team.
Can we define retention and deletion rules?
Retention requirements can be aligned with the selected product plan and deployment. Dedicated and customer-managed environments support more specific operational rules, which are agreed during solution design.
Is a data processing agreement available?
Yes. hypescale GmbH provides data processing terms for customers and documents the relevant service providers in the processing chain.
What does the EU AI Act mean for customer-service chatbots?
The AI Act’s chatbot transparency obligations apply from 2 August 2026. Users must be informed that they are interacting with AI. Assess the duties for your use case and how that information is presented. EU hosting alone does not establish AI Act compliance.
Turn your requirements into a deployable architecture
Bring your security questionnaire, data residency rules, and target workflows. We will map the right deployment and model path with you.