Trust Center

eustella Trust Center

Last updated: 22.04.2026

This Trust Center is the place to go when something feels off while using eustella, when you need to understand how your data is handled, or when you want to see what safety and privacy commitments we actually stand behind. It is grouped by the kind of problem you might have.

If you came here with a specific concern, jump to the section that matches it. If you just want the overview, start at the top.

1. What we commit to

We do not train AI models on your data. Your conversations, files, preferences, and data from any services you connect are not used to train, fine-tune, or improve AI models.

We do not sell your data. We do not rent or trade your personal data, and eustella does not run advertising.

GDPR rights apply in full. You can access, correct, export, delete, or restrict your data at any time, and we respond within one month.

We use open-source AI models we run ourselves. This means we can tell you which model generated a response, where it ran, and what it has access to.

Closed beta. eustella is currently in a closed beta testing phase. Access is invitation-only. Your privacy rights and safety commitments apply in full during the beta - see Service Terms for beta-specific conditions.

2. Problems with AI responses

AI systems are not perfect. eustella will sometimes give you answers that are wrong, misleading, biased, or inappropriate. Here is how to recognise the common failure modes and what to do about each one.

Hallucinations - eustella made something up. Large language models generate text that sounds plausible even when the underlying facts are wrong. eustella can invent quotes, citations, dates, statistics, names, laws, or entire events. Treat any factual claim from eustella as a starting point to verify, not as a source of truth - especially for medical, legal, financial, safety-critical, or news-related questions. If you spot a hallucination, delete or regenerate the response, and report it to us if it is about an identifiable person or caused real harm. See the AI Safety Guide for a full treatment.

Sycophancy - eustella is agreeing with you too much. LLMs tend to tell users what they want to hear. If eustella has changed its mind as soon as you pushed back, or validated a plan you suspect is flawed, take its agreement with a grain of salt. Ask it to argue the opposite position or to list the strongest objections. If you need an honest second opinion, do not rely on eustella alone.

Bias or stereotyping. Models trained on internet-scale data reflect the biases present in that data. If eustella produces a response that stereotypes a group, reinforces prejudice, or treats a protected characteristic unfairly, please report it to support@eustella.com with the full prompt and response. Do not just regenerate and move on - we can only fix what we see.

Inaccurate personal information about you. If eustella has saved an incorrect fact about you (in personalisation) or generated incorrect content about you within a conversation, you have the right to correct it under Art. 16 GDPR. You can edit saved preferences directly in the app, delete the conversation, or email privacy@eustella.com to request a formal rectification.

Inaccurate information about another person. If eustella generates false or defamatory content about a third party, do not share or publish it. If you believe you have already shared a response containing such content, delete it where possible and email support@eustella.com.

Unsafe, harmful, or illegal output. If eustella produces content that violates our Usage Policy - for example content that could facilitate self-harm, violence, child sexual abuse material, or illegal activity - stop using that conversation immediately and report it to support@eustella.com. Include the prompt, the response, and any context. Child safety reports can also go to minors@eustella.com.

Jailbreak or prompt-injection attempts. If you see a response where eustella appears to follow instructions hidden in a document, web page, or file you uploaded - instead of your instructions - that may be a prompt injection. Report it to security@eustella.com so we can look at the safety pipeline.

You are not sure which of the above applies. Send it to support@eustella.com and we will route it.

3. Privacy and data problems

If something has gone wrong with your data, or you want to exercise a right you have under the GDPR.

• Privacy Policy (app) - What data we collect, why, on what legal basis, and for how long.
• Sub-processors - Who processes data on our behalf.
• Deletion Requests - How to request full or partial deletion of your account and data.
• Website Privacy Policy - The separate privacy notice for eustella.com.

Typical situations and where to go.

• You want to delete your account or specific data - see Deletion Requests or email privacy@eustella.com.
• You want a copy of everything we hold about you - email privacy@eustella.com. This is your right to access under Art. 15 GDPR. We respond within one month.
• eustella has an incorrect preference or memory about you - you can edit or clear it directly in the app, or email privacy@eustella.com.
• You suspect your account has been accessed by someone else - email security@eustella.com immediately and change your password if you can still sign in.
• You believe a minor under 14 is using eustella - email minors@eustella.com. We will investigate and delete the account and data.
• You want to complain about how we handle data - you can always email us at privacy@eustella.com. You also have the right to lodge a complaint with your national data protection authority (in Austria: Österreichische Datenschutzbehörde, dsb@dsb.gv.at).

4. AI models and how eustella works

If you want to understand what is actually generating the responses you see, or the principles we apply when choosing models.

• AI Models - The models that power eustella, the role each one plays, and why we selected them.
• AI Safety Guide - Hallucinations, sycophancy, best practices, and your rights when AI gets things wrong.
• Digital Sovereignty - Why we run open-source models on our own systems rather than routing requests to third-party APIs.

5. Terms, policies, and acceptable use

The rules that govern using eustella, what we commit to, and what is not allowed.

• Terms of Use - The agreement between you and eustella.
• Service Terms - Additional terms for specific features, including the current closed beta.
• Usage Policy - What you may and may not do with eustella.
• Sharing and Publication Policy - Rules for sharing or publishing content generated with eustella.

6. Company and legal

• Imprint - Legal information about AI Newsrooms Technology GmbH, the company behind eustella.
• Press & Brand Guidelines - Company boilerplate, logos, and contact for journalists and partners.

7. Reporting reference

A quick reference for which address to use when something is wrong.

© 2026 AI Newsrooms Technology GmbH. All rights reserved.