Immersee

About the Project

Immersee is a Telegram mini app with an AI companion for an adult audience.

Users could create a virtual character, customize their appearance, buy additional items, use in-app currency, and interact with the character through chat and content generation.

The main product experience happened inside the Telegram chat: conversation, roleplay, image generation, and other AI-based scenarios. My area of responsibility was the mini app itself — the interface layer where users managed the character, settings, purchases, and internal resources.

It was important to make the product feel emotional and atmospheric without making it explicit or vulgar. The interface had to support a sense of mystery, privacy, and premium quality.

Task

At the start, the team already had a clear understanding of the feature set and technical constraints. The product was built around Telegram, an AI backend, and an existing interaction logic, so the main task was not to invent new mechanics, but to quickly wrap the product into a coherent interface.

I had to design the complete visual UI for the mini app: all key screens, states, purchase flows, character settings, in-app currency logic, and the final handoff for development.

I worked closely with a product manager. The functional scope was mostly defined from the beginning, so my role was focused on the visual system, interface logic, and preparing the product for implementation.

Visual Direction

The project did not have ready-made visual references or a strict design system. The team’s expectation was mostly emotional: the interface needed to feel premium, unusual, thematic, and slightly provocative.

I proposed a dark visual direction with soft lighting, noisy textures, glass-like surfaces, and bright accent buttons. This helped create a more private, evening-like, slightly futuristic atmosphere.

The main challenge was balance. The content itself was already quite provocative, so the interface could not become too aggressive. It had to support the character rather than compete with it, turning the whole experience into a calm, controlled, and coherent system.

Store and Character Customization

The store was the most complex section. It had to combine a visual catalog, item cards, prices, purchase states, selected item preview, and the flow for topping up currency.

I started with this section because it best represented the future logic of the whole product: cards, modal states, large images, purchase actions, confirmation, insufficient balance, and return to the main scenario.

For the store interface, I used large item previews, soft rounded corners, a dark background, and high-contrast price tags. This made it easy for users to understand what was available, what had already been purchased, and how much each item cost.

After selecting an item, the user moved to a detailed preview screen. There, they could view the item in a larger format, see the price, understand whether they had enough currency, and either purchase the item or top up their balance.

Interface Approach

Because of the nature of the visual content, the interface had to remain relatively restrained. I did not want to amplify the provocative side of the product with excessive decoration, aggressive typography, or overloaded graphics.

So I used a modular approach:

1. I started with the most complex screens and states using the store as the main reference point.

2. Then I gradually built the other sections based on that logic.

3. After each stage, I refined the previous screens based on feedback.

4. At the end, I prepared the complete design for development handoff.

This process helped move quickly while keeping the product visually consistent.

Monetization and In-App Currency

The product used two main resource mechanics: energy and crystals.

Energy was connected to content generation. Crystals were used to buy items and unlock additional features. For these scenarios, I designed separate screens for purchasing resources, selecting packages, continuing the flow, and receiving referral rewards.

These screens had to be clear and direct: users needed to quickly understand which resource they needed, how much they would get, and what action they should take next.

Outcome

As a result, I prepared the complete UI design for the mini app: key user scenarios, visual system, store, character settings, internal resource screens, referral program, and development-ready states.

The product was launched in beta with my design implemented. After that, my involvement in the project ended. The bot is currently no longer active, but the project remained an interesting experience for me at the intersection of an AI product, a Telegram mini app, a virtual character, and an adult audience.

The main challenge here was not the number of screens, but the tone: to make the interface emotional, atmospheric, and commercial, while keeping it from becoming overloaded or visually crude