tanuo.ai operates a turnkey, "plug-and-play" AI software stack — that is, a coordinated bundle of individual software and infrastructure building blocks that together run as one ready-made platform — for small and medium-sized enterprises, GDPR-aligned and on dedicated infrastructure in the EU. The stack combines four application building blocks: a chat interface, workflow automation, document processing and a knowledge base. It is based on open-weight AI models that run on dedicated hardware — without dependency on individual US providers.

The core of our offering is the operation and automation of this stack (the "tanuo Connector"), not the sale of software. You receive a ready-to-use platform that we set up, operate and maintain for you - without having to build your own AI team.

Yes. Via open standards (object storage/S3, OpenAI-compatible API, OpenAPI, workflow automation), existing systems such as DMS, ERP or CRM can be integrated. The integration of AI tools via MCP is in development.

Your data is stored with tenant isolation per customer instance in the EU (Germany) and, by default, is not shared with external AI providers. Key protection mechanisms:

• Zero-trust network: Access and internal communication run over an encrypted VPN overlay with fine-grained access control.

• Centralized identity / single sign-on via dedicated identity management

• Secrets management via a dedicated, encrypted manager.

• AI gateway as a control point: All AI requests run through a single, controllable gateway.

To offer an immediately usable platform, tanuo.ai operates its own AI inference hardware. In the entry-level packages, this inference hardware is shared (marked with "*") — comparable to shared server hardware on which the customer instances nevertheless remain cleanly separated from one another. So only the computing power at the infrastructure level is shared, not your data or your instance. To ensure all customers receive fair performance, a fair-use policyapplies: under very high, sustained load, speed is throttled fairly instead of blocking access. Those who use dedicated inference hardware (BYOK, higher packages or their own hardware) are not subject to any such limitation.

tanuo.ai relies on powerful open-weight modelsoperated on dedicated hardware. Depending on the package, additional model types are available, such as image/language understanding (vision models) or speech recognition (speech-to-text). The model selection is tailored to the chosen package.

Optionally, and only on explicit request, external AI providers can be connected via your own access key (BYOK, OpenAI-compatible). By default, no no external model APIs are used — requests do not leave your instance toward external AI providers.

No. tanuo.ai is a managed service provider. We set up the AI stack on your behalf, automate it and keep it running. You remain the operator and licensee of the open-source components used; you pay for the managed operation (hosting, automation, maintenance), not for the software as such. Access by Tanuo is, as a rule, exclusively controlled via automation — for updates and adjustments.