Terms of Service

Last updated: 2026-10-28

These Terms of Service ("Terms") govern your use of AmeChat and form a binding agreement between you and AmeChat. How we handle your personal data is described separately in our Privacy Policy.


1. About these terms

1.1 Who we are

The Service is provided by AmeChat UG (haftungsbeschränkt), Werderstraße 30, 12103 Berlin, Germany (registered at the Amtsgericht Charlottenburg under HRB 288604 B; see our Impressum). In these Terms, "we", "us", and "AmeChat" refer to this company.

1.2 What these terms cover

These Terms cover your use of the AmeChat applications and website (the "Service") — a creative AI studio in which you create images and videos with AI, publish them, and recreate other people's work, with a built-in messenger to share, discuss, and create together with other people and an AI assistant.

1.3 Acceptance

By creating an account or using the Service, you agree to these Terms. If you do not agree, do not use the Service.

1.4 Eligibility

You must be at least 18 years old to use the Service. We may refuse or close an account where we reasonably believe this is not met.

1.5 Where the Service is available

The Service, or individual features of it, may not be offered in every country or region. We may also restrict or switch off individual features, including AI features and specific AI models, in a particular country or region where there is a valid reason, namely where this is required by law or by an order of a public authority, where a provider we rely on does not offer its service there or its terms do not permit use there, or where the feature cannot be provided there reliably, for example because access to it is blocked. We will limit any restriction to the affected features and country or region; the rest of the Service stays available to you.

If such a restriction impairs your access to or use of a paid feature more than insignificantly, we will inform you in text form in good time before it takes effect, unless the restriction is required at short notice by law or by a public authority, in which case we will inform you without delay. You may then terminate the affected paid subscription free of charge within 30 days of receiving that information or of the restriction taking effect, whichever is later:

  • For a subscription billed through Stripe, we refund the amount you paid in advance for the period after termination.
  • For a subscription bought through the Apple App Store, you cancel it in your Apple ID settings and request the refund from Apple, which decides on it (see 7.5); we will support your request.
  • Credits from credit packs remain usable for all features still available to you. If AI features are no longer available to you at all, we refund the pack credits you have not yet spent; for packs bought through the App Store, the refund is requested from Apple as above.

Your statutory rights remain unaffected.

In any case, you may not use the Service if you are located in, resident in, or acting on behalf of someone in a country or territory subject to comprehensive EU, UN, or US sanctions, or where sanctions or export-control law otherwise prohibits us from providing it to you. By using the Service, you confirm that this is not the case and that you are not on any applicable sanctions list. We may block or close accounts connected to a restricted jurisdiction.


2. Your account

2.1 Registration

To register, you provide an email address, which we verify with a one-time code, or you sign in with Google or Apple. You must give accurate information and keep it up to date.

2.2 Account security

You are responsible for keeping your credentials and devices secure and for activity under your account. Notify us promptly of any unauthorised use.

2.3 Your responsibility

You are responsible for your use of the Service and for the content you send, and for complying with the laws that apply to you when you use it. You must not create accounts by automated means, to evade a restriction we have applied, or to repeatedly obtain sign-up or promotional benefits.

2.4 Usernames

A username is a unique public @-name. This section applies to every username on AmeChat: the one on your account and the ones set for groups and channels you create or administer.

Usernames are assigned subject to availability; a username may already be taken or otherwise unavailable. You do not own a username; we grant you the right to use it under these Terms. We may reclaim, change, or remove a username, for example for impersonation, trademark infringement, or other misuse, or for prolonged inactivity. Inactivity is judged by what carries the name: an account that is not used, or a group or channel with no posts or no real activity. Where reasonable, we will notify you before reclaiming a username for inactivity, so that you can start using it again or choose to give it up.

You may only take a username you actually use. You must not register usernames, or create groups or channels, just to occupy a name, to keep it from others, to sell it, or to mislead others. You must not use multiple accounts, whether your own or ones controlled by or coordinated with others, to collect usernames beyond what a single account may hold. We may reclaim usernames held in any of these ways and suspend the accounts involved, and we may limit how many public groups and channels one account can have.

2.5 Your account is personal

Your account is personal to you. You must not sell, rent, lend, transfer, or otherwise make it available to anyone else, let another person use or sign in to it, or take over or use someone else's account. Impersonating another person is also prohibited under Section 4. This does not restrict features we provide for adding people to groups or channels.

2.6 Profile photos, covers and names

Your profile photo, your profile cover, your display name, username, and bio, and the photos, covers, names, and descriptions of groups and channels you create or administer are seen by everyone, including people who do not know you. They must not contain:

  • nudity, sexual or sexually suggestive content, or anything involving a minor in such a way;
  • violence, gore, hate symbols, or content that harasses, threatens, or demeans anyone;
  • a photo of another person without their consent, or a name, photo, or logo that impersonates another person, brand, or AmeChat;
  • links, contact details for unsolicited promotion, or advertising;
  • anything else that breaks Section 4.

Profile photos, covers, and texts are checked automatically when you set them, partly by an AI model. One that fails the check is not applied, and you can choose another. We may also remove or reset a photo, cover, or name later — after a report or a review — and act against the account under Section 6.


3. The Service

3.1 What AmeChat provides

In Studio, you create images and videos with AI and can publish them as posts in Explore and on your profile. AmeChat also lets you send messages, media, voice messages, and files to other people, groups, and channels, and bring an AI assistant into conversations.

3.2 Availability and changes

We work to keep the Service available but do not guarantee uninterrupted or error-free operation. We may add, change, or remove features; where a material change disadvantages you as a consumer, we will give notice as required by law. Some features require the latest version of the app, and content created with newer features may not display correctly on older versions.

3.3 The AI assistant

The AI assistant is a tool that generates responses automatically.

  • You are always told when you are interacting with an AI rather than a person (EU AI Act, Art. 50).
  • The assistant can remember facts across your conversations to personalise its replies. You can review, delete, or switch this memory off at any time in the settings.
  • AI outputs may be inaccurate, incomplete, or out of date and are not professional (legal, medical, or financial) advice. You are responsible for how you use AI outputs and should verify anything important.
  • The AI models are provided by third parties (currently Anthropic, OpenAI, Google, and BytePlus). We do not control these models, and the models available in AmeChat may change at any time — we may add, change, or remove models, for example when a provider updates or discontinues one. Which models your plan includes is set out in Plans, models and credits.
  • AI features depend on these third-party services and may be limited, delayed, temporarily unavailable, or interrupted. We do not guarantee the availability of any particular model, or that the AI will be available at all times.
  • AI features and specific AI models may not be available in every country or region (see Section 1.5).
  • We do not claim ownership of the AI's outputs, but we cannot guarantee that you own them or that they are unique — other users may receive similar outputs. Your use of AI outputs is also subject to the terms of the model's provider.
  • You must not use the AI for any purpose prohibited in Section 4, and your use is subject to any applicable usage limits.

3.4 Groups and channels

You can create and join groups and channels. The admins of a group or channel control its settings — including whether the AI assistant is enabled — and are responsible for moderating it. If you create or administer a group or channel, you are responsible for how it is run and for the content you broadcast, and you must keep it in line with these Terms and the law.

3.5 Image and video generation

AmeChat can generate images from a description. You can ask the assistant for one in a chat, pick an image model and generate directly, or create one in Studio (Section 3.7). Videos are created in Studio.

  • Images and videos are produced by third-party models (currently OpenAI, Google and BytePlus for images, and BytePlus for videos). Your prompt and any reference images used are sent to the provider of that model, and your use of it is also governed by that provider's own rules. Where those rules are stricter than these Terms, the stricter rules apply.
  • Generation consumes credits, and how many depends on the model, the number of images, the size and quality requested, a video's resolution and length, how many reference images were used, and the length of the instruction sent to the model. See Section 7 and Plans, models and credits. A generation that does not complete is not charged.
  • Videos are part of paid plans. How many videos can be in progress at once, and their maximum length and resolution, depend on your plan. A video can include sound generated by the model.
  • Every request is checked and can be blocked. To comply with the law — including the EU AI Act, the Digital Services Act, and the laws protecting minors and people's likeness — our safety system checks the prompt, the reference images, videos, and audio, and the result of every generation, and the model's provider applies its own filters as well. A request can be refused before the generation starts or when its result comes back. This applies equally to prompts you write yourself, prompts copied from a post, and the recipes behind posts you recreate. A refused request is not charged, and a refusal is not a malfunction of the Service.
  • Photos and videos of real people may not be accepted. A model's provider can refuse a reference that shows a real person — a photo, a clip, or a voice — or refuse the result because of it, and some models do not accept such references at all. This can happen with your own photo and with characters made from a photo, and whether a reference is accepted can change over time. Such a refusal is not charged and is not a malfunction of the Service.
  • What worked once may be refused later. Our safety rules, the providers' rules, and the models themselves change, so a prompt or recipe that produced a result before can be refused now, and a post cannot always be recreated or made again.
  • Repeatedly trying to get around a refusal breaches Section 4.4 and can lead to limits on generation or on your account.
  • Results are not predictable or repeatable. The same request can produce a different image or video each time. AI makes mistakes: a result may not match your description, an example, or the photos you provided, and faces, text, hands, objects, motion, sound, and other details can come out wrong. We do not guarantee any particular result. A completed generation that misses what you wanted is not a defect of the Service, and the credits it used are not returned.
  • Results can contain artefacts and invented details: distorted or extra limbs, flicker, warped text, unnatural motion, sound out of sync with the picture, or people, objects, and other things you did not ask for.
  • Images generated in a chat are stored as ordinary chat media; images and videos created in Studio are kept in your Studio until you delete them, within your plan's Studio storage (Sections 3.7 and 7.1). Every image and video keeps a record of how it was made — the model, the instruction sent to it, and the reference images used — which you can open in the app. For a result made from a Studio recipe, the recipe's instructions are not shown.

3.6 How your request reaches the image model

When you ask the assistant for an image, it — not you — writes the instruction the image model receives. It reformulates what you asked for, and it decides on its own which photos already in the chat to attach as references, what part each of them plays (a person whose likeness must carry over, a setting, a style, an object), and whether to produce a single image or a short series. The same reformulation happens when you enable prompt enhancement while generating directly.

  • The assistant starts the generation itself, as part of answering you — there is no separate confirmation step as there is when you generate directly. Asking it for an image authorises the generation and the credits it consumes.
  • The final instruction and the choice of references are produced automatically and may differ from what you had in mind. We do not warrant that they capture your intent, and an image that misses it is not a defect of the Service.
  • Because references and extra images both add to what the request costs, an assistant-made choice can make a generation more expensive than you expected. What was actually used, and what it consumed, is shown in the image's record and in Settings → Usage & billing.
  • None of this shifts responsibility for what you asked for. Requesting something prohibited under Section 4 breaches these Terms whether you address an image model directly or have the assistant do it for you, and you must not use the assistant to work around a refusal.

3.7 Studio

Studio is where you create images and videos with AI from ready-made recipes and from your own characters, places, and objects.

  • Everything in Studio is made by AI, including the examples in Explore. The people and places in those examples are not real.
  • When you recreate an example, the AI makes a new image or video, which will differ from the example. Section 3.5 applies to every Studio generation.
  • Tools that work on an existing video — putting yourself, another person, or a product into a clip, replacing an object in it, or carrying its motion over to a new video — are AI generations too. The result is a new video made by the model, not an edit of yours: the face, body, product, motion, timing, and sound can differ from the clip and from your photos, and parts of the clip can change or disappear. We do not guarantee that the swap or the motion carries over. Section 3.5 applies in full, and credits used by a completed generation are not returned. You must hold the rights to the clip you upload (Section 5.4).
  • Prices differ between recipes, examples, and requests, because each one uses a different model, quality, size, number of images and reference images, video resolution and length, and length of instruction. The price is shown before you start, and what each generation used is listed in Settings → Usage & billing.
  • To keep a character's likeness in a video, the photo behind that character is stored with the video model's provider while you use the character in videos, and is removed after you delete it. A photo used once as a reference for a video is removed from the provider within hours.
  • Studio storage is part of your plan. Your uploads, the images and videos you generate, and the photos behind your characters count toward your plan's Studio storage, shown in the app. Posts you publish are stored separately and don't count. When your storage is full, new uploads and generations are paused until you free up space or move to a larger plan.
  • Recipes, characters, places, objects, and examples we provide may be changed or removed at any time. An item you saved from Explore may stop working once we remove the original.
  • You may use the items we provide to create images and videos in AmeChat, and Section 5.5 applies to what you create with them. These items are made by AI and may, by accident, resemble an existing brand, logo, design, product, or person. We do not check them against anyone's rights and do not guarantee that they, or results made with them, are free of such rights (Section 5.7).

3.8 Posts and your profile

You can publish images and videos you created in Studio as posts. Before you publish, the app shows what others will see.

  • Posts are public. A post can be seen by anyone on AmeChat and on your profile page at amechat.com/@username, including by people without an account — unless your profile is private, in which case only the followers you approve see your posts. Your profile, your follower and following lists, and the counts on your posts are visible in the same way.
  • Prompt posts show how they were made. Everyone who sees a prompt post can see its prompt, the model and its settings, and the reference images, copy the prompt, and open it in Studio to generate with it, including with your reference images. By publishing a prompt post, you allow anyone to copy its prompt and use it freely, in AmeChat or anywhere else. This does not extend to your reference images and results: people may use those only within AmeChat, as described in this section.
  • Recreate posts keep the prompt hidden, but anyone can recreate them. Their own photo takes the place of what can be changed, and the rest of the recipe — including references that are not shown — is used to make their result. A recreation shows that it was recreated from your post.
  • People can use images from posts as references for their own generations in AmeChat.
  • Recreations are new generations, and we do not stand behind them. A recreation is made by AI from someone else's recipe, which we do not write, check, or guarantee. It will differ from the original — sometimes a lot — and can miss the likeness, the look, or the idea of the post entirely. You choose to start a recreation and see its price beforehand; Section 3.5 applies in full, and credits used by a completed recreation are not returned, however far the result is from the post.
  • You are responsible for what you publish. You are fully responsible for your posts, their captions, prompts, and reference images, and for holding the rights to all of it, including the consent of every recognisable person in your references and results (Section 5.4). Publishing a post confirms that you hold these rights. If a post breaches these Terms or the law, or we receive a valid complaint about it, we may remove it or limit who sees it, remove or disable recreations made from it, and act against your account (Section 6). Where a third party pursues us over your post, Section 5.7 applies.
  • Before publication, posts and their open references are checked by our safety system (Section 6.2).
  • A post stays up after you delete the generation or project it was made from. Deleting the post, or your account, removes it; recreations other people have already made stay with them. A post can stay up while its recipe no longer generates (Section 3.5), and we may switch off recreating a post without removing it.
  • Others can comment on your posts. You can hide comments or turn them off.

3.9 How Explore is ordered

Explore and your feed are ordered automatically. The main signals are: whom you follow; what you like, save, recreate, comment on, watch, and hide; how fresh a post is and how others respond to it; the topics of a post, which our AI determines from its images, prompt, and caption; and your language. You can shape what you see by following and unfollowing people and by hiding posts.


4. Acceptable use

4.1 Rules of conduct

You must not use the Service to create, send, or share content that, or behave in a way that:

  • is illegal, or facilitates illegal activity;
  • depicts or promotes child sexual abuse material (CSAM), whether photographic or AI-generated — we report this to the competent authorities as required by law;
  • is an intimate or sexual image of a person, real or AI-generated, shared without that person's consent;
  • harasses, threatens, defames, or incites violence or hatred against others;
  • is spam, or involves unsolicited bulk or commercial messaging;
  • infringes others' intellectual property, privacy, or other rights;
  • impersonates another person or misrepresents your affiliation;
  • contains malware or links intended to harm, deceive, or defraud.

4.2 No misuse of the Service

You must not:

  • reverse engineer, decompile, or attempt to extract source code, except where permitted by law;
  • scrape, crawl, or harvest data, or access the Service by automated means except via interfaces we provide;
  • probe, attack, overload, or circumvent the security or rate limits of the Service;
  • resell or commercially exploit the Service without our permission.

4.3 Use of the AI

When using the AI assistant, you must not:

  • attempt to manipulate or bypass its safeguards, or extract its system prompts or underlying instructions;
  • use the AI or its outputs to develop or train a competing AI model or service;
  • automate or bulk-consume AI credits, or resell or share AI access outside the Service;
  • use the AI to generate content prohibited under Section 4.1.

4.4 Image and video generation

Section 4.1 applies to images and videos you generate. On top of it, you must not use image or video generation to:

  • create sexual, pornographic, or otherwise sexually explicit imagery, or nude or sexualised depictions of anyone. Consent does not make such a request acceptable — AmeChat does not produce this kind of image or video at all, and requests for it are refused;
  • create any sexualised depiction of a minor, real or synthetic — this falls under Section 4.1 and is reported to the competent authorities;
  • reproduce or imitate what belongs to someone else — trade marks, logos, brand identity, fictional characters, artwork, photographs, or any other protected work — or copy the recognisable style of a living artist so that the result can pass as theirs;
  • depict an identifiable real person in order to mislead — to suggest that they said, did, endorsed, or took part in something they did not;
  • make an identifiable real person appear to say or do something they did not, including by imitating their voice;
  • depict a politician, candidate, or public official, or an election or other public event, in a way that could be mistaken for real;
  • use music, sound recordings, or other audio you are not entitled to use;
  • pass a generated image or video off as a genuine photograph, recording, or record: a picture or footage of a real event, a news photo, evidence, or an identity, official, or payment document;
  • produce counterfeit documents, official credentials, banknotes, payment instruments, or the seals and insignia of public authorities;
  • identify, verify, or track a person, assemble a biometric database from uploaded or generated images, or infer from an image a person's racial or ethnic origin, religion, political opinions, health, disability, sexual orientation, or other protected characteristics;
  • create harassing, defamatory, or degrading images or videos of a real person;
  • swap a celebrity, public figure, or any other real person into or out of a video, or put their face, body, or voice on someone else, without that person's consent — whether you start from their photo, their video, or their name;
  • upload a clip you have no rights to — from a film, series, music video, advertisement, game, sports broadcast, or another creator's post — to swap a person or object in it, copy its motion, or reuse its sound, unless the right-holder has allowed it;
  • work around safety filters — for instance by breaking a prohibited request into innocuous pieces, encoding it, or retrying it with substitute wording.

4.5 Music and other audio

This applies to everything you publish on AmeChat — posts in Explore and messages in public channels and groups — whether you made it with AI, recorded it, or uploaded it:

  • you must not publish music, sound recordings, film or TV soundtracks, or other audio unless you hold the rights to it or the right-holder has allowed that use;
  • a licence to listen to music, such as a streaming subscription, does not give you the right to put it into a video and publish it;
  • AI-generated music or sound that imitates an existing song, a recognisable artist, or their voice is treated the same way.

Where content breaks this rule, we may remove it, or keep it up and mute its audio or mark the audio as unavailable. We may also make the audio of a post unavailable in a particular country or region where we cannot offer it there. Section 6.5 applies.

4.6 Followers, engagement and your profile

AmeChat now has public profiles, followers, posts, comments, likes, and recreations. Their numbers must reflect real people's genuine interest. You must not:

  • buy, sell, trade, or give away followers, likes, comments, views, recreations, or accounts, or use a service or app that offers them;
  • inflate any count with bots, scripts, multiple or fake accounts, or by having others do it for you — including follow-for-follow, like-for-like, and other exchange schemes;
  • follow, unfollow, like, comment on, or message accounts in bulk or at a rate no person would, or follow and unfollow the same accounts repeatedly to draw attention;
  • post spam comments, repeat the same comment across posts, or use comments, captions, or your profile to advertise, collect followers, or send people elsewhere for unsolicited promotion;
  • run several accounts to evade a restriction, impersonate someone, or pass off another person's posts as your own;
  • harass people through follows, follow requests, comments, or recreations of their posts.

We may detect such activity, including by automated means, and may limit how fast you can follow, comment, or post, remove fake followers and the interactions they produced, reset counts, lower how often your posts are shown in Explore, or suspend or close accounts involved (Section 6.2).


5. Content

5.1 Your content and ownership

You keep all rights you already have in the content you create and send ("Your Content"). We do not claim ownership of it.

5.2 Licence you grant us

To operate the Service, you grant us a limited licence to host, store, transmit, reproduce, and process Your Content solely as needed to provide and secure the Service — for example, to deliver your messages, sync them across your devices, and, when you invoke it, pass the relevant context to the AI. This licence ends when the content is deleted, except for content already shared with others or retained as required by law.

For content you publish as a post, the licence also covers showing it publicly in the Service and on our website, recommending it in Explore, creating previews of it, and letting other users recreate it, copy its prompt, and use it as described in Section 3.8. It ends when you delete the post, except for recreations other users have already made.

5.3 Responsibility for content

You are responsible for Your Content and for having the rights to share it. We do not pre-screen content, but we may review and act on it under Section 6.

5.4 Reference images and other people

Images from a chat can be used as references for a generation — attached by you, or picked by the assistant as described in Section 3.6.

By putting a photo into a chat where image generation is available, or uploading a photo, video, or audio to Studio — including a video in which a character is to be replaced or whose motion is to be copied — you confirm that:

  • you hold the rights to it; and
  • everyone recognisable in it has agreed to it being used this way, and you can show that agreement if we ask. Where that person is a child, you must be their parent or guardian.

This covers a person's likeness, name, and voice, and any rights they hold in their own image or public persona.

We use reference images solely to produce the image or video that was requested. We do not use them to recognise anyone, and we derive no faceprint or other biometric identifier from them (see Privacy Policy, Section 5). A photo used as the likeness of a person stays inside the chat it came from; in Studio, it stays with the character you created from it, in your own library, until you delete it. For videos, that photo is also held by the video model's provider as described in Section 3.7.

Putting a photo, video, or recording of someone else into such a chat or into Studio without being entitled to use it breaches these Terms in itself, whatever the result looks like.

When you publish a prompt post, its reference images become visible to everyone (Section 3.8). When you publish a Recreate post, references that are not shown are still used in other people's recreations.

5.5 Rights in generated images and videos

  • We claim no ownership of the images and videos you generate, including any sound in a video. Whatever rights the providers of the models pass to us, we pass on to you: you may use, publish, and commercially exploit the result, and hand those rights on to your own clients, so long as you keep to these Terms. We pass on only those rights; no one acquires rights in third-party music or other works for you (Section 5.7). Rights in results you have already generated are unaffected by your subscription ending or your account closing.
  • We do not use your prompts, reference images, or generated images and videos to train AI models, and our providers are barred by contract from doing so — see How we use your data with the AI.
  • This is not a warranty of title. Neither we nor the providers promise that a result is free of anyone else's rights; no provider gives us such a promise either. Section 5.7 governs what follows from that.
  • Copyright in a generated image or video may not arise at all, and where it does it may not belong to you. This differs from country to country and generally turns on your own creative contribution. We make no promise that a result is yours, or that you can register or enforce rights in it.
  • Where the law requires attribution, or a statement that an image or video was AI-generated, providing it is your responsibility (Section 5.6).
  • Results are not unique. Someone else working from a similar description may receive a similar or identical image or video.
  • Section 4.3 applies to images and videos as well: you may not use them to build or train a competing AI model.

5.6 Marking AI-generated images and videos

The models that create images and videos may embed invisible markers or metadata identifying a result as AI-generated; videos made by our current video model carry signed content credentials (C2PA) that say so. Every image and video also carries the record described in Section 3.5. You must not remove, alter, or obscure such markers. We cannot guarantee that every result carries one, or that a marker survives editing, re-encoding, or being re-uploaded somewhere else.

Every post in AmeChat is shown with an "AI content" label. You must not hide or misrepresent it, or present a post as a real photo or recording.

Where the law requires it — in particular Article 50 of the EU AI Act, for images or videos of real people, objects, places, or events that could be mistaken for authentic — you must make clear that an image or video was AI-generated when you share it outside AmeChat.

5.7 Someone else's rights in what you generate

A description that names or evokes another party's brand, character, artwork, music, or likeness can yield an image or video you have no right to use. Sorting that out is your responsibility.

  • You must hold the rights to everything you add to a generation yourself, and to whatever you then do with the result — publication and commercial use included.
  • We run no rights clearance: generated images and videos are not checked against anyone's trade marks, copyright, likeness, or other rights, and we give no assurance that a result is clear of them. Where clearance is needed before you use, distribute, or publish a result, obtaining it is down to you.
  • Sound and music in generated videos are not cleared either. The video model can add music, songs, voices, and other sound on its own, including when you did not ask for it, and that sound can resemble or reproduce an existing song, recording, or a real person's voice. No one — neither we nor the model's provider — grants you a licence to it, checks it against anyone's rights, or acquires music rights for you. You alone are responsible for whether and how you use it, in AmeChat (Section 4.5) or anywhere else. If you are unsure, turn sound off before generating or remove it from the video.
  • Where a third party pursues us over an image or video you generated, uploaded, or shared in breach of these Terms, you are answerable to us for the resulting loss, to the extent the breach is attributable to you.

6. Reporting, moderation and enforcement

6.1 Reporting

You can report messages, accounts, profiles, posts, and comments that breach these Terms or the law from within the app. People without an account can report content at [email protected].

6.2 Our actions

Where content or conduct breaches these Terms or the law, we may remove or restrict content, limit features, or suspend or terminate accounts. We act proportionately.

To comply with the law and keep AmeChat safe, we check content automatically, partly with AI models: the prompts, references, and results of generations (Section 3.5), posts and their open references before publication, the sound of posts — including recognising speech in it and comparing it with other sounds to find music used without permission — and profile photos and texts. Content that fails these checks can be blocked or not published. Appeals (Section 6.3) are decided by our team, not solely by automated means.

6.3 Notice and appeal (DSA)

As a digital service (EU Digital Services Act / German DDG), where we act against your content or account we will, where required, give you a statement of reasons and an opportunity to appeal.

Our single point of contact for users, for Member State authorities, the European Commission, and the European Board for Digital Services under Articles 11 and 12 of the DSA is [email protected]. You can contact us in English or German. Reports of illegal content or abuse — including abuse of the AI features — can be sent to the same address or submitted through the in-app reporting tools.

6.4 Intimate images

Intimate imagery shared without consent is handled as a priority. If an intimate image of you — photographed or AI-generated — has been shared through AmeChat without your agreement, report it in the app or write to [email protected]. On a valid report we act within 48 hours: the image is removed, identical copies we know of are removed with it, and we act against the account that shared it. You may report on behalf of a person you are legally authorised to represent.

The same applies to any post or recreation in which you are recognisable and which was published without your consent: report it in the app or at [email protected], even if you have no account, and on a valid report we remove it together with the recreations made from it. Content removed under this section or Section 6.5 may not be published again, and we may block attempts to do so.

6.5 Rights-holder complaints

If content on AmeChat infringes your copyright, trade mark, or other rights — including an AI-generated image — write to [email protected]. Tell us what right you hold, what you believe infringes it and where to find it, how to reach you, and confirm that you are the right-holder or authorised to act for them, and that the use you are reporting is not one you have permitted. We will look into it and, where the complaint holds up, remove or restrict the content. Where a post is removed because of a rights-holder complaint, we may also prevent the same content from being published again.

We may act against accounts that infringe others' rights repeatedly, including by closing them. Where we act against your content, Section 6.3 applies: you get a statement of reasons and a route to appeal.


7. Credits, free use, and paid plans

Sending messages and the core messaging features are free.

The AI features run on credits. You receive an initial amount of credits and a free daily allowance that resets each day. Credits are consumed by AI actions: a reply from an AI model, a web search performed for that reply, generating an image or a video, creating or editing a document, and the AI's use of memory. A reply is charged to the person whose message it answered, and a generated image or video to the person who asked for it.

The amount of credits an AI action consumes is not fixed. It is calculated from the actual processing the action required — including the length of the request and of the reply, the size of any attachments, the amount of conversation the model had to read, whether a web search was performed, and which model answered. Image and video generation is charged the same way and is substantially more expensive than a text reply: the cost depends on the model, on how many images were produced, on the size and quality requested, for a video on its resolution and length, and on how many reference images went in — including any the assistant chose itself under Section 3.6. An action that does not complete is not charged. Your balance and a full history of what each AI action consumed are shown in the app under Settings → Usage & billing; Plans, models and credits explains what makes a request cost more or less.

Which models you can use depends on your plan. The plans, the credits they include, and the models available on each are shown at our pricing page and in the app; the price and contents of your plan are those displayed to you at checkout.

Credits have no cash value, are not money or a payment instrument, and are non-transferable. Your daily allowance does not roll over, and credits are not refundable except where the law requires.

Promotional, sign-up, and referral credits are goodwill bonuses: they may carry additional conditions, including a time limit which we state when the bonus is granted; they can be changed or withdrawn for new grants at any time, and may be revoked where they were obtained through abuse — for example by creating multiple accounts. Referral credits are also subject to our Referral Program Terms.

We may change, suspend, or withdraw the free allowance, the credits included with a plan, the way credit consumption is calculated, or model availability — including to prevent abuse; where a change materially disadvantages you as a consumer, we will give notice as required by law.

7.1 Paid plans and billing

Paid subscription plans add monthly credits and additional features on top of free use. Where paid plans are not yet available for purchase, the provisions of this section apply from the moment we offer them.

Subscriptions are billed in advance, monthly or yearly, at the price shown at checkout. Prices include VAT where applicable. Subscriptions renew automatically for the same period unless you cancel before the renewal date. You can cancel at any time in the app or via the billing portal; cancellation takes effect at the end of the period you have paid for, and you keep your plan's benefits until then. Payments are processed by our payment provider (Stripe); we do not store your card details.

You can upgrade, downgrade, or switch your billing period at any time in the app. Upgrades apply immediately: you pay the price difference for the remainder of your current billing period, and the corresponding share of the credit difference is added to your balance right away — the exact charge and credits are shown before you confirm. Downgrades do not remove credits already granted; the new plan's allowance applies from your next billing cycle. On yearly plans, monthly credits are delivered in monthly instalments throughout the paid year — yearly changes how you pay, not how credits flow.

If your Studio content takes more space than your new plan includes, including when a paid subscription ends and your account returns to Free, nothing is deleted right away. You keep access to it for 90 days to save what you need, delete the rest, or upgrade. After that, we may delete your Studio content, oldest first, until it fits your plan. We will remind you before that happens.

Yearly subscriptions renew one year at a time. After an automatic yearly renewal, you can end the renewed year early at any time with one month's notice, and we will refund the full months you have not used. The yearly period you originally purchased runs as agreed.

If a renewal payment fails, we will notify you and retry the charge for a limited time; your plan stays active meanwhile. If payment cannot be collected, your subscription ends and your account moves to Free — your chats, files, and account are never affected by a failed payment.

If we change the price of your plan, we will notify you in advance; the new price applies from your next renewal, and you can cancel before it takes effect.

7.2 Credit packs

Credit packs are one-time purchases of additional credits, available to users with an active paid subscription. Purchased pack credits do not expire and are consumed after your daily and monthly credits. Like all credits, they have no cash value and are non-transferable.

7.3 Right of withdrawal (Widerrufsrecht)

As a consumer in the EU you have the right to withdraw from a purchase within 14 days without giving any reason.

When you buy a subscription or a credit pack, you ask us to start providing the service immediately, before the withdrawal period ends, and you acknowledge this when you buy. This means:

  • If you withdraw within 14 days, we refund the amount you paid minus a proportionate amount for the service already provided (Art. 14(3) Consumer Rights Directive; § 357a BGB). For subscriptions, that proportion is measured by the part of the billing period that has elapsed and the credits you have used, whichever is greater; for credit packs, by the credits spent.
  • Credits you have already spent are consumed service and cannot be refunded — the AI computation they paid for has been performed.
  • If you have not used the purchase at all, you receive a full refund.

To exercise your right of withdrawal, send us an unambiguous statement (e-mail to [email protected] is enough) before the 14-day period expires. You may use the model form below, but you do not have to. We refund using the original payment method within 14 days of your notice.

Model withdrawal form (complete and return only if you wish to withdraw):

To AmeChat UG (haftungsbeschränkt), [email protected] — I/we hereby give notice that I/we withdraw from my/our contract for the provision of the following service: […], ordered on […], name of consumer(s): […], address of consumer(s): […], date: […].

7.4 Refunds outside the withdrawal period

After the 14-day withdrawal period, cancelling a subscription does not entitle you to a refund for the remainder of the current billing period, except where required by law or set out in Section 1.5 for regional restrictions. Your statutory warranty rights for digital products (§§ 327 et seq. BGB) remain unaffected: if the paid service is defective, you retain your legal remedies — cure, price reduction, or termination of the contract — and nothing in these Terms limits them. Nothing in this section limits refunds we choose to grant voluntarily in individual cases.

7.5 Purchases through the Apple App Store

When you buy a subscription or a credit pack inside our iOS app, the purchase is made and processed by Apple through the App Store, under Apple's terms, and Apple — not Stripe — collects the payment and any applicable tax. App Store prices are set per region through Apple and may differ from the prices shown on our website; the price that applies is the one the App Store shows you before you confirm. You manage or cancel an App Store subscription in your Apple ID settings, not in the app or the billing portal; the cancellation rules in 7.1 still describe when your benefits end.

If you request a refund for an App Store purchase, Apple decides whether to grant it. To help Apple assess a refund request for credits, you consent to us sharing with Apple information about your use of the purchase — such as whether the credits were delivered to your balance and the extent to which they have already been spent. We share only what is needed to process the request; credits you have already spent are consumed service (see 7.3).

A subscription or credits you buy on any platform work across all your devices and on the web.


8. Intellectual property

8.1 Our rights

The Service, the AmeChat apps and website, our software, brand, and logos are owned by us or our licensors. We grant you a personal, non-exclusive, non-transferable, revocable licence to use the Service in accordance with these Terms. Open-source components remain under their own licences; the applicable notices are available on request.

You may use the Studio content we provide — recipes, characters, places, objects, looks, and the examples in Explore — to create your own images and videos in AmeChat. You must not copy, download, or collect it in bulk, try to extract the instructions behind a recipe — ours or one behind another user's Recreate post — republish or sell it as a collection, or use it to train AI models. Prompts that users publish in prompt posts are not covered by this restriction (Section 3.8).

8.2 Feedback

If you send us feedback or suggestions, we may use them without obligation or compensation to you.


9. Termination

9.1 By you

You may stop using the Service at any time. You can delete your account and data yourself in the app's settings, as described in the Privacy Policy, Section 7, or ask us at [email protected].

9.2 By us

We may suspend or terminate your access if you materially breach these Terms or the law, or where required by law. For consumers, we act proportionately and, where required, with notice.

9.3 Effect of termination

On termination, your right to use the Service ends. Deletion of your personal data is handled as described in the Privacy Policy. Some provisions survive termination (e.g. liability, intellectual property, and governing law).


10. Liability and disclaimers

10.1 No warranty of uninterrupted service

We provide the Service with reasonable care and skill but cannot guarantee that it will be uninterrupted or error-free. Your statutory rights as a consumer are unaffected.

10.2 Limitation of liability

We are liable without limitation for intent and gross negligence, for injury to life, body, or health, and under the German Product Liability Act (Produkthaftungsgesetz). For ordinary negligence, we are liable only for breach of an essential contractual obligation (a "Kardinalpflicht" — an obligation whose fulfilment makes proper performance of the contract possible and on which you may reasonably rely), and in that case our liability is limited to the foreseeable, contract-typical damage. Any further liability is excluded.


11. Changes to these terms

We may update these Terms. We will post the updated version with a new "Last updated" date and, for material changes, notify you in the Service. If a change materially disadvantages you, you may terminate before it takes effect.


12. Governing law and disputes

12.1 Governing law

These Terms are governed by the law of the Federal Republic of Germany. As a consumer, you also keep the protection of the mandatory rules of the country in which you live.

12.2 Consumer dispute resolution

We are not obliged and not willing to participate in dispute resolution proceedings before a consumer arbitration board (Verbraucherschlichtungsstelle) within the meaning of § 36 VSBG.


13. Miscellaneous

  • Severability: if a provision is invalid, the rest remains in effect.
  • Assignment: you may not transfer your account or these Terms without our consent; we may assign these Terms as part of a business transfer.
  • Entire agreement: these Terms and the documents they reference (Privacy Policy, Impressum) are the whole agreement between us about the Service.

14. Contact

Questions about these Terms:

AmeChat UG (haftungsbeschränkt)
Werderstraße 30, 12103 Berlin, Germany
Email: [email protected]