Independent review for Ireland · public sources · 19.10.2026
The public presentation is emotional and cinematic. The terms are more concrete: Eva AI describes a virtual-friend chatbot using text, voice and, where available, video calls. This guide reads both surfaces together.
The short answer
Eva AI is an 18+ virtual-friend app whose published iOS terms describe text and voice chat, video calls where available, and synthetic AI-generated content. Access is divided between Basic, Premium and Neurons, but benefits and prices can vary by region and subscription length, so the live checkout matters more than an undated price quoted elsewhere. ’s privacy policy says text and voice messages may be used to improve the service and processed by external providers; treat a companion chat as personal data, not as a private diary. Eva AI
Start with the product evidence
SEEN
≠
HUMAN
01 / What Eva AI is
Eva AI presents itself as a companion application built around ongoing dialogue. The public home uses the line “Feel seen” and says conversations should click from the first message. The public About page expands that positioning: characters are described as engaged, attentive and curious, with responses that are intended to feel natural. It also says they remember details and open up as the conversation continues. Those statements explain the intended experience, but they should not be read as proof of human awareness or perfect memory.
The Android terms define the “Virtual friend” as an artificially intelligent simulation. They describe communication through text, voice and, where available, a video-call interface. The same terms say AI-generated content is synthetic and may not represent real people, events or interactions. That legal definition matters because the visual language deliberately borrows cues from dating, attention and closeness. The product may imitate a responsive relationship, yet the underlying interaction remains generated software.
The terms also make clear that Eva AI is not medical care, mental-health care or another professional service, and should not be used in an emergency. A companion may provide conversational structure or entertainment, but a convincing tone is not a clinical qualification. People seeking urgent or professional help should use a suitable human service instead.
02 / How it works
The current Android terms state that account holders must be at least 18. They also explain that account information must be accurate and current. This age requirement comes from Eva AI’s own terms, not from the imagery used in this guide.
The terms say users can create a virtual character by choosing or uploading relevant content. They require users to hold the necessary rights and permissions for anything uploaded, including images, and prohibit using another person’s likeness without consent.
Text and voice are explicitly named in the terms. Video calls are described as available only where supported, and may require camera and microphone permissions. The feature can be changed, suspended or discontinued, so availability should be checked in the current app.
The terms describe Basic and Premium programmes as well as an in-app unit called Neurons. They say the exact Premium benefits appear during subscription and prices may vary by region and term. This guide therefore does not publish a fixed price or pretend every feature is included.
The terms do not guarantee accuracy, realism or suitability of AI-generated content. Treat character replies as synthetic entertainment and illustration. Check important facts elsewhere, and never assume a natural-sounding response is verified or professionally reviewed.
03 / Public screenshots
Every image below was captured from evaapp.ai or evaai.ai/brandbook in a clean public browser session. The files were resized for the web, not composited or generated. No OurDream media was required, so no character, model or feature from another service is attributed to Eva AI.
04 / Reading the brandbook
The brandbook relies on very large, thin uppercase lettering. It makes short messages feel deliberate and premium, and it leaves enough empty space for each statement to land. This guide references that scale without reproducing the official page structure.
Black backgrounds dominate, while violet and electric blue appear as concentrated light rather than constant decoration. The palette supports intimacy and technology at the same time. It is more restrained than a typical neon chatbot interface.
The public brandbook explicitly links contrast to logo readability. That is why this page keeps body copy bright on dark surfaces and avoids placing the brand name over noisy screenshots. Readability is treated as a functional constraint, not just an aesthetic preference.
The brandbook recommends the EVA AI lettering on white when a partner already uses its own symbol. This independent guide does not claim partnership or official status. It uses a separate wordmark and identifies itself as an external publication.
05 / Practical checks
Confirm whether voice, video calls, images and character creation are available on your device, account and current programme. “Where available” is a real limitation in the published terms.
Video calls may use the camera and microphone. Review device permissions and disable access you do not want to grant. Permission is a practical choice, not a requirement to trust every generated response.
Only upload material you have the right to use. Do not upload another person’s image without consent, and do not ask a synthetic character to imitate a real person in a misleading way.
Read the subscription screen before purchase. The terms describe Basic, Premium and Neurons, but say benefits and prices can vary. Check renewal, cancellation and marketplace rules in the actual checkout.
A conversational tone can encourage disclosure. Keep passwords, payment details, identity documents, workplace secrets, private health records and other people’s information outside the chat.
Do not use the app as emergency, medical or mental-health care. The service’s own terms reject that role. Contact a qualified professional or emergency service when the situation requires one.
Eva AI’s public material is unusually useful because the emotional home page, concise About page, detailed terms, privacy policy and formal brandbook reveal different layers of the same product. The home explains the desired feeling. The About page describes conversational behaviour. The terms define the service, age threshold, communication modes, synthetic nature, payment structure and limits. The privacy policy explains that chat and usage data can be processed to run and improve the service. The brandbook explains how the identity should remain recognisable. Read together, they support a grounded description: is an adult virtual-character companion service built around synthetic conversation, with text and voice plus video calls where available. They do not support a promise of human understanding, flawless memory, guaranteed therapeutic benefit, absolute confidentiality or a universal fixed price. Eva AI
Common Ireland questions
No. Eva AI’s terms define the virtual friend as an artificial-intelligence simulation and say generated content is synthetic. The same terms say the service is not medical care, mental-health care or an emergency service. A warm or convincing reply should therefore be treated as generated conversation, not human judgement or professional advice.
The current iOS terms describe a limited Basic programme, a paid Premium programme and Neurons used for some interactions or purchases. They also say Premium benefits are defined during subscription and that prices can vary by region and term. This guide does not publish a fixed Ireland price: review the current App Store or in-app checkout, renewal period and cancellation controls before paying.
Text and voice are named service modes, while video calls are described as available only in some cases. The terms say video calls may require camera and microphone permissions and may be changed or withdrawn. Check the permission prompt on your own device and deny access that is unnecessary for the way you intend to use the app.
Do not assume so. The privacy policy says text and voice messages can be collected to improve the virtual friend and may be processed with external services. It also describes usage data, analytics and optional session recordings subject to consent. Read the current policy, use deletion or access rights where applicable, and keep passwords, financial records and third-party secrets out of chat.
No. The terms say users are responsible for uploaded content and must have the necessary rights. They prohibit uploading images of other people without consent and prohibit deceptive likeness or impersonation uses. Use only material you own or are authorised to use, and check whether the character is marked private or public in the current interface.
Proceed with the limits in view