# Waterglass Waves: Managed AI Agents from Europe > Waterglass Waves is the European AI agent platform: define AI agents yourself, we run them on EU infrastructure – with freely interchangeable AI models. Canonical: https://www.waterglass.ai/en/products/waves/ Language: English (Deutsch: https://www.waterglass.ai/de/produkte/waves.md) Company: Waterglass FlexCo, Börseplatz 1/3/6, 1010 Vienna, Austria · hi@waterglass.ai ## Waterglass Waves: The European AI agent platform With Waves your team defines AI agents itself – we run them on European infrastructure. Freely interchangeable AI models, full control over data, access and costs. Your AI agents in Waterglass Waves ## European AI agents: Create and manage your AI agents With Waterglass Waves you can create and manage custom AI agents – the platform runs them on European infrastructure. Every AI agent runs in an isolated environment, with schedules, persistent memory and a record of every action. You are free to choose which AI model an agent uses, and you can switch it anytime – including to models we run in Europe ourselves. Waves is built by Waterglass: we create and operate [AI agents for businesses](https://www.waterglass.ai/en/ai-agents/). With Waterglass Waves we put the tools in your hands to build your own AI agents and run them on European infrastructure. Your AI Model ## System Architecture: The elements of an AI agent *Trigger* - Schedule - Events - Requests - Email - Chat - Phone *Managed AI Agent* *Decisions and actions* - Makes decisions - Sends replies - Updates systems - Creates documents - Hands off to people Hosted in Europe by Waterglass *Brain* Understands the tasks, works on them, and picks the right AI model for the job: *Knowledge* Draws on your company's documents and data to work on tasks in the right context. *Memory* Remembers the current task and earlier context to continuously learn new insights. *Tools* Connects to your existing apps and systems to interact with them: *Approval* The AI agent follows clear rules and requires a person to sign off on anything sensitive. *Logs* Everything the AI agent does is logged, so that all actions are documented. ## For business customers: Operated on European infrastructure ### Data center in Nuremberg Waves runs in a data center in Nuremberg, so compute, object storage and data residency are always within the EU. ### Isolated environment per AI agent Every session runs in its own isolated environment. You decide whether network access is possible via allow-lists. ### Encrypted credentials If you like, your AI agents can also use credentials. These are stored securely encrypted. ### Control and traceability Optionally use two-factor authentication and org-scoped API keys for added security. ## How to start: 3 steps to your AI agent Waves is in early access: we open up access gradually and onboard every team personally. ### Get access Enter your email address and we get back to you shortly with the next steps. ### Personal onboarding We set up your organization and define the first agents together with your team – aligned with your workflows and privacy requirements. ### Adjust and extend yourself From then on you use the AI agent platform yourself, with all its features. ## Questions & answers: Common questions about Waves **What are managed AI agents?** Managed AI agents are agents you only define, not operate – the platform takes care of sandboxing, schedules, memory and records. The concept was popularized by offerings such as Anthropic's managed agents, which run on US infrastructure with a single model family. Waterglass Waves brings the concept to Europe: free choice of AI model, EU data residency, and versioned agent definitions that belong to you. **Where is my data processed?** The platform runs in a data center in Nuremberg, including object storage – data residency is in the EU. For AI models, Waves transparently shows each provider's jurisdiction, and you decide which model an agent uses. **Which AI models can I use?** Currently models by Anthropic (Claude), OpenAI (GPT), z.ai (GLM), Moonshot (Kimi) and Mistral. The model is chosen per agent and can be overridden per session. Models hosted in Europe by Waterglass are coming soon. **Can I adjust my AI agents myself?** Yes. Your AI agents are versioned definitions you edit in the console as YAML or through the REST API. During onboarding we set up the first AI agents together. After that your team works independently, and we stay available. **How safe are credentials and keys?** Credentials live in vaults and are encrypted (AES-256-GCM). Values are write-only and never appear in any response or log. Console access is protected by two-factor authentication, and API keys are shown once and stored hashed. **Can Waves be used in a GDPR-compliant way?** Your data is processed on European infrastructure, in isolated execution with allow-lists and an event record. We discuss your specific requirements – such as which models and data flows are permissible – during onboarding. **What does Waves cost?** There is no public price list during early access. Billing is usage-based and transparent: you see tokens and costs per day plus a monthly projection in the console. We sort out the details during onboarding. **When can I start?** Enter your email address and we get back to you within 1–2 business days. ## Get your access to Waves Sign up and we reach out personally with a demo and answers to your questions. ### How can we help? Book a short call and we show you Waves live and figure out how it fits your business.