LEGAL
Terms of Service
Note: This is a convenience translation. The German version (AGB) is legally authoritative — see clause 12.2.
Terms of Service for KindChat.
End-to-end encrypted · On-device AI protection · GDPR- & DSA-compliant · For users aged 16 and over.
Last updated: June 2026 (Art. 50 AI Act applies from 2 August 2026).
Contracting party: sonrisa ventures GmbH, Kleiststraße 35, 10787 Berlin.
1. Scope and Conclusion of Contract
1.1 These Terms ("Terms") apply to all users of the KindChat app (iOS/Android) and the associated services ("Service" or "KindChat"), operated by sonrisa ventures GmbH ("provider", "we"), based in Germany.
1.2 The Service is directed exclusively at consumers resident in the EU (and, where applicable, further supported countries). By registering / first use, you agree to these Terms.
1.3 The contract is concluded through download, installation and successful registration (phone number + confirmation). App-store terms apply in addition.
1.4 Not permitted for users under 16 years of age. By registering, you confirm that you are at least 16 years old or have reached the minimum age applicable in your country of residence.
2. Description of Service — What KindChat Is and Is Not
2.1 KindChat is an end-to-end encrypted messenger (Signal protocol with post-quantum cryptography) for text, voice, video and media messages as well as calls, with integrated protection features against harmful content (insults, hate, threats, fraud, phishing, unwanted/harmful images/videos/voice messages).
2.2 The protection features serve exclusively to protect users; KindChat is not a surveillance tool; the provider does not read content and does not create profiles / advertising analyses.
2.3 Key characteristics: moderation primarily on-device (before encryption/sending) using automated procedures including AI; in rare cases of doubt, additional purpose-bound checking on EU servers (clause 3); Protection Circle (optional, up to 6 people, consent-based, without access to content); Pulse feed only for your own contacts (no algorithm, no strangers); no advertising / no tracking / no data sales; no guarantee of 100%ige detection (clause 3.2).
3. AI-Supported Content Moderation and Transparency (Art. 50 AI Act)
3.1 Automated content checking (Art. 50 para. 1): KindChat uses automated procedures including AI to detect harmful content and protect you (insults, hate, threats, fraud, phishing, unsolicited/harmful images, videos, voice messages). Checking takes place as far as possible directly on your device, before encryption/sending. In cases of doubt, additional checking on EU infrastructure. E2EE remains unaffected; content is not analysed for advertising/profiling purposes.
3.2 Limits of moderation (limitation of liability — central clause): No guarantee of completeness. No automated system detects harmful content with complete certainty. Protection mechanisms substantially reduce the risk but cannot eliminate it. False-negative and false-positive results are possible. KindChat gives no warranty that all harmful content is detected, blocked or obscured. It does not replace your own caution. You retain control: view, report, block the sender.
3.3 Human oversight (Art. 14 AI Act in conjunction with the DSA): Automated decisions can be erroneous. Right to review via the report/complaint function (DSA); in suitable cases, review by a natural person. Safety-critical cases are not decided conclusively by automation alone.
3.4 Care features and consent (Protection Circle): Features that support a person with management/setup (co-sharing, remote setup) only with explicit consent that can be revoked at any time. Not a surveillance tool; no one gains insight into content; actions taken on behalf of a protected person are visible to that person.
3.5 Age requirement: from 16 years of age; by registering you confirm the minimum age. KindChat is not a service for children.
3.6 Linguistic limits: detection quality may vary by language; no warranty of equivalent detection in all languages.
3.7 No generative AI / no labelling obligation (Art. 50 para. 2): The moderation AI does not generate content, it only classifies; therefore no watermarking obligation. Future generative features would be made transparent and labelled separately.
4. User Obligations and Prohibited Use
4.1 Use only for lawful purposes, in accordance with these Terms and applicable law.
4.2 Prohibited in particular: sending/distributing illegal content (e.g. CSAM, extremist content, fraud, threats, insults within the meaning of the German Criminal Code (StGB)); circumventing/manipulating the protection features; harming third parties or the provider; multiple registration / account sharing; reverse engineering, decompilation, unauthorised access.
4.3 In the event of a breach: removal of content, temporary/permanent suspension, termination without notice (in compliance with the DSA requirements on Notice & Action and complaint procedures).
5. End-to-End Encryption and Server-Side Escalation
5.1 Continuous E2EE; the provider generally has no access to the plaintext content of messages, calls, media.
5.2 In strictly limited exceptional cases (high probability of harmful content or a legitimate user report) server-side escalation to EU infrastructure: content is briefly and purpose-bound decrypted and checked; thereafter deleted without delay. Serves exclusively to protect; described in the privacy policy.
5.3 Does not undermine the E2EE promise for normal use; comparable to "break-glass" in the event of acute danger; transparently communicated.
6. Liability of the Provider
6.1 Unlimited liability for injury to life, body, health as well as for intent/gross negligence.
6.2 Otherwise liability only for breach of essential contractual obligations (cardinal obligations), limited to the typically foreseeable damage.
6.3 No liability for undetected harmful content. The limits from clause 3.2 apply without restriction; no guarantee of completeness/freedom from error of the automated protection features; liability for indirect damage, consequential damage, lost profit — to the extent permitted — is excluded.
6.4 The limitations of liability also apply in favour of legal representatives, employees, vicarious agents.
7. Data Protection
Processing in accordance with the privacy policy, an integral part of these Terms.
8. Termination, Suspension and DSA Complaint Procedure
8.1 Terminable at any time by deleting the app and the account (in the settings); data is deleted in accordance with the privacy policy.
8.2 Suspension/termination in the event of serious/repeated breaches; before permanent suspension — as far as possible and legally required (DSA) — reasons and an opportunity to comment.
8.3 DSA complaint procedure (Art. 20 DSA): against moderation decisions (block, blur, warning) a complaint via the in-app report function; review within the statutory deadlines; right to review by a natural person (clause 3.3).
9. Changes to These Terms
9.1 Adjustment where necessary; notification with a reasonable period (at least 14 days) by email or in-app.
9.2 No objection within the period = acceptance; in the event of objection, termination without notice is possible.
10. Applicable Law and Place of Jurisdiction
10.1 Law of the Federal Republic of Germany, excluding the UN Convention on Contracts for the International Sale of Goods (CISG).
10.2 For consumers resident in the EU, the mandatory consumer-protection provisions of the country of residence apply; place of jurisdiction for consumers: place of residence or — at the provider's choice — Berlin.
10.3 For all other disputes, the exclusive place of jurisdiction is Berlin.
11. Consumer Rights and the OS Platform
11.1 Statutory right of withdrawal; for digital services it lapses upon commencement of performance (access/first use) with express consent.
11.2 EU platform for online dispute resolution (OS): https://ec.europa.eu/consumers/odr. We are neither obliged nor willing to participate in proceedings before a consumer arbitration board.
12. Final Provisions
12.1 Severability clause: the invalidity of individual provisions leaves the rest of the contract unaffected; an invalid clause is replaced by a valid one that comes closest to its economic purpose.
12.2 These Terms are authoritative in the German language; translations serve information purposes only.
12.3 Contact for legal questions: hi@kindchat.app.
Appendix / references: privacy policy, AI transparency under Art. 50 AI Act.