Zentre gives teams one place to use the world’s leading AI models — with every byte staying in the EU. Multiple models, shared knowledge, intelligent agents, and complete governance. All hosted on European infrastructure.
Organizations in the EU/DACH region that need governed AI — from non-regulated businesses, SMEs, construction, and the skilled trades (Handwerk) to regulated and confidentiality-sensitive sectors such as law firms, tax advisors, healthcare, and financial institutions.
Yes. The service is directed exclusively at businesses within the meaning of § 14 of the German Civil Code (BGB).
No feature limitations. You’ll get full access to Zentre with up to €5 in AI credits and a maximum of 5 users. Once the credits are exhausted, you’ll need to upgrade to a paid plan.
Super Admin — Full workspace access and control.
Admin — Manage users, teams, models, agents, and knowledge bases.
Billing Admin — Access billing, invoices, and usage information.
Member — Access chat, agents, and assigned knowledge bases.
Yes. Zentre supports the latest models from leading AI providers, all delivered through European infrastructure.
Yes. Zentre supports document creation and editing. This feature is currently in Beta.
Zentre supports PDFs, images, Markdown files, Word documents, text files, and other common business formats.
Yes. You can switch between supported models without losing conversation context.
Chat is for everyday conversations, research, and content creation.
Agents are pre-configured assistants with instructions, knowledge, and workflows tailored to specific tasks or teams.
Send your request to support@zentre.ai. We’re continuously expanding our integration library based on customer demand.
All customer data is hosted on European infrastructure and remains within the EU.
No. Your conversations, documents, and knowledge bases are never used to train public AI models.
Zentre knows your organisation. By combining AI models, organisational knowledge, agents, and governance, Zentre delivers AI that works with your business context.
A range of models routed through an EU-based gateway, including EU providers (e.g., Mistral), EU-hosted open models, and — where you choose — EU endpoints of established model providers. You can restrict which models are available (see the AI Models and Professional Secrecy sections).
EZTO TECHNOLOGIES GmbH, Am Brand 41, 55116 Mainz, Germany (AG Mainz HRB 54349).
Zentre is operated in line with the GDPR and the German Federal Data Protection Act (BDSG). For customer data processed in the platform, EZTO acts as a processor under Art. 28 GDPR.
Both, depending on context: a processor for the data you process in Zentre (under the DPA), and an independent controller for website, contract, and billing data (under our Privacy Policy).
Yes. An Art. 28 GDPR DPA is available, including annexes for the processing description, technical and organizational measures (TOM), subprocessors, and retention.
As a processor, we support you (the controller) in fulfilling data subject rights, including export and deletion functions and reasonable assistance. Direct requests we receive are forwarded to you rather than answered independently.
No. Zentre does not use customer content to train its own models, and training is excluded with the integrated model providers where contractually agreed (e.g., via zero data retention).
With Scaleway (Scaleway SAS, France), region Paris (fr-par); productive data is processed within the EU/EEA.
Yes. Hosting and AI processing run within the EU/EEA, and we do not use AI models hosted outside the EU.
Your AI chat content and usage data are processed within the EU/EEA and are not transferred to third countries. The only ancillary services that may involve a provider outside the EU are email delivery (Infomaniak, Switzerland) and payment processing (Stripe); these operate under appropriate safeguards (adequacy decision and/or EU Standard Contractual Clauses).
Differing EU-hosting, on-prem, or Private Cloud options are available for Enterprise by agreement.
TLS 1.3 for data in transit and AES-256-GCM for data and artifacts at rest.
We operate an information security management system (ISMS) aligned to ISO/IEC 27001. ISO/IEC 27001 certification is underway.
Through logical tenant separation. For professional-secrecy use (§ 203 mode), a dedicated, isolated tenant with its own knowledge base is provided.
Yes, we support MFA via SSO.
Requests are routed by default through the EU-based AI gateway Cortecs (Cortecs GmbH, Vienna), which processes within the EU and engages the downstream model providers as its own subprocessors.
Yes — zero data retention is available, so prompts/outputs are not written to persistent storage at the gateway, subject to the configuration and the individual provider.
No. A no-training commitment applies to API usage, and is passed down to the integrated providers where contractually available.
If enabled, search queries are sent to the EU search provider Linkup (France, with a zero-data-retention option). Web search is deactivated by default in § 203 mode.
Yes. Model selection can be restricted via an allowlist; in § 203 mode it is limited to authorized EU/EU-controlled endpoints.
Yes, via the § 203 mode together with a separate confidentiality agreement. A standard Art. 28 DPA alone is not sufficient for § 203 — a dedicated confidentiality undertaking with instruction (Belehrung) is required.
A configuration for processing protected secrets: a dedicated, isolated tenant (own knowledge base), EU hosting, EU/EU-controlled models, zero data retention, and web search disabled.
Tier A (EU-controlled): only EU-controlled or EU-self-hosted models — no provider with a non-EU parent. Tier B (EU-hosted): additionally EU endpoints of established (incl. US-vendor) models, with zero data retention and SCCs; the customer acknowledges the residual exposure (e.g., US CLOUD Act).
Yes — a signable confidentiality agreement covering Germany (§ 203 StGB), Austria (§ 121 öStGB), and Switzerland (Art. 321 CH-StGB), including the statutory instruction and a flow-down to subprocessors. Profession-specific variants (BRAO, StBerG, healthcare) and the statutory wording annex are available.
The EU gateway Cortecs has confirmed it will enter an additional § 203 confidentiality undertaking, including the Austrian (§ 121 öStGB) version.
No. The DPA covers data protection; professional secrecy additionally requires the separate confidentiality agreement described above.
Yes. A DORA addendum is available for financial entities and ICT third-party service providers and applies on conclusion of the main agreement.
The service is designed to be used in a manner where it does not support critical or important functions, so EZTO is not a critical ICT third-party service provider; if that assessment changes, a separate arrangement under Art. 30(3) DORA is negotiated.
Notification without undue delay and assistance for ICT incidents relating to the services EZTO provides directly; for third-party (gateway/model/search) incidents, available information is passed on.
Hosting: Scaleway (EU). AI gateway: Cortecs (EU); model providers are engaged via Cortecs. Web search: Linkup (EU). Plus email (Infomaniak), billing (Stripe), website consent (Usercentrics) and analytics (Plausible). The current list is published in the Trust Center.
At least 30 days before a change takes effect, via the Trust Center / subprocessor list.
Yes — a data flow document describing data categories, recipients, locations, legal bases, and retention is available in the Trust Center.
The General Terms and Conditions (B2B), the DPA, and — where applicable — the § 203 confidentiality agreement and/or DORA addendum.
German law; place of jurisdiction is Mainz, Germany.
The German version is legally binding; English versions are provided for information purposes only.
Electronically. The DPA and DORA addendum are incorporated on conclusion of the main agreement; the § 203 confidentiality agreement is concluded as a separate, signable document.
You can cancel through an official email or via the admin console.
Monthly plans run for one month and renew automatically by one further month unless you cancel by the end of the current billing month; the cancellation takes effect at the end of that month.
Annual plans run for twelve months and renew automatically for a further twelve months unless you cancel before the end of the current term; the cancellation takes effect at the end of the term.
Fees for the current (already-started or prepaid) period are not refunded pro-rata; the right to extraordinary termination for good cause remains unaffected. For unused prepaid credits, see the Pricing & Credits section.
A single seat-based plan (per user), plus Enterprise and Private Cloud options by agreement.
Credits are prepaid usage units for AI consumption. A service fee of 7.5% applies to credit purchases and is non-refundable.
Credits remain usable while the account is actively used. If the account is unused for 36 consecutive months, unused credits expire; we send reminders (~60 and ~14 days) beforehand, and any login or use resets the period.
On termination: amounts of €50 or more are refunded automatically to the last-used payment method; amounts below €50 on request within 60 days. If EZTO ends the contract without cause or discontinues the service, unused credits are refunded automatically regardless of amount.
Yes, for Enterprise customers by agreement; in that case the prepaid expiry/refund rules do not apply.
In the standard EU setup and in Tier A (EU-controlled models), there is no exposure to US-vendor access. Exposure can arise only if you select EU endpoints of US-vendor models (Tier B), which is acknowledged on selection.
