← Back to Work

Nemora — a mobile-first AI
companion, designed to convert

Product and growth design for nemora.ai — a shipped, revenue-generating AI companion platform. Not a concept: every screen here is in production, and every design decision answers to conversion, retention and paying users. (18+ product — imagery curated for portfolio.)

Product & Growth Designer
Live in production
Mobile-first web
Product · Ads · SEO · Campaigns
LiveReal users, real revenue
30+Handoff flows (desktop + mobile)
100+Ad creatives designed
9Hero variants A/B tested
Mobile-first — most traffic on phones
nemora.ai — live
Nemora Discover
Loads the real nemora.ai in this frame
Nemora mobile home
Real mobile site

Not a screenshot — the actual product, live. Tap either frame to load nemora.ai. Mobile-first: the phone frame is the primary experience.

Context: designing where every pixel answers to revenue

Nemora is an AI companion platform — users create a character, chat, and generate images. It's a competitive, conversion-driven market: traffic is bought with ads, and the product has seconds to turn a visitor into a registered user and days to turn them into a subscriber. My work covered the full loop: product UI, the paywall system, pricing, and the entire growth layer — ad creatives, SEO landing pages, blog templates, seasonal campaigns.

It's also an 18+ product, which raises the design bar in a specific way: users want discretion and trust as much as delight. That shaped decisions from billing copy to how imagery is teased.

Competitor analysis

The AI-companion market is crowded and paid-traffic-driven. I benchmarked the products users compare Nemora against before committing paid ad spend against them:

Nemora
Character.AI
Replika
Candy.ai
Janitor AI
Mobile-first, one-tap onboarding
~
~
Deep character creator
~
In-chat image generation
~
Discreet / anonymous billing
~
Polished, trustworthy UI
~

← swipe the matrix →

The pattern: the big players (Character.AI, Replika) are trusted but tame; the adult-native tools (Janitor, Crushon) are permissive but feel unsafe and unpolished. Nemora's opening was a product that looks as trustworthy as the mainstream apps while being as open as the adult ones — with discretion engineered in.

Who I designed for

M
Marcus, 26Curious first-timer · arrives from an ad, on his phone
Phone only From paid ad Wants discretion

I clicked out of curiosity. If it takes more than a few taps or feels sketchy, I'm gone.

Intent to pay
Patience
Privacy concern
Goals
  • Try it instantly, no friction
  • See value before paying
  • Zero trace on his phone or bank
Frustrations
  • Forced signup before any value
  • Sketchy-looking checkout
  • Obvious charges on a statement
D
Devin, 31Engaged returning user · creates and collects characters
Daily, mobile Power user Subscriber

I've made five characters. I want them to feel distinct — and I want new things to keep coming.

Intent to pay
Patience
Privacy concern
Goals
  • Rich, controllable character creation
  • Fresh content and reasons to return
  • Manage several companions easily
Frustrations
  • Characters that feel same-y
  • Hitting limits mid-conversation
  • No sense of progress or reward

Marcus drove the frictionless mobile entry, the tease-don't-block paywall, and the discreet billing; Devin drove the depth of the character creator, the My-AI management view, and the retention systems (achievements, seasonal drops). The whole funnel is designed to walk a Marcus into a Devin.

Journey map — ad click to subscriber

Land
Explore
Engage
Hit the wall
Subscribe
Doing
Taps an ad, lands on Discover on his phone
Browses characters, opens one, starts chatting free
Sends messages, gets a blurred image teaser
Wants the image / more messages — sees the paywall
Picks 12-month, pays via discreet checkout
Feeling
skeptical
curious
hooked
tempted, wary
reassured
Risk
Bounces if it looks cheap or asks to sign up first
Leaves if characters feel generic
Drops if the free experience feels crippled
Abandons if the paywall feels greedy or unsafe
Cart-abandons at an untrustworthy checkout
Design response
Polished Discover, no signup gate to browse
Personality tags + hover-to-chat on every card
Full chat rhythm free; premium is teased, not blocked
Benefit-matched paywall + "no adult transaction" line
Stripe badges, card logos, cancel-anytime, anonymity

← swipe the map →

User flows

User flow — first session (free → paid)
Ad click Discover (no gate) Pick a path Open a character Create your own Chat free Blur paywall Upgrade? SubscribePremium unlocked Keep chatting freeRe-teased later
User flow — create a character
Create My AI Ethnicity & look — visual presets Personality & voice Occupation, hobbies, style Generate Success — meet your AI Start chatting

The pricing page: a conversion system, not a table

This screen carries the business. Every element is a deliberate conversion mechanic:

nemora.ai/subscription
Nemora subscription page with countdown, three anchored tiers, premium benefits and trust signals
The subscription page — countdown, anchored tiers, benefit ladder, and a trust stack aimed at this product's specific anxiety.

Anchoring & urgency

Three tiers priced to make 12 months the obvious deal ($13.99 → $5.99 with the crossed-out $25.99 anchor), a first-visit discount countdown, and "only today" framing. Aggressive — but every claim is real and cancelable.

Trust for a discreet product

The line that converts here isn't the price — it's "No adult transaction in your bank statement." Plus Stripe-badged checkout, card logos, cancel-anytime, and "100% anonymous" up top. Discretion is the feature.

The paywall that doesn't feel like a wall

Premium content in chat isn't hidden — it's teased. Images arrive blurred with an inline "Upgrade to Premium" chip, so free users experience the full conversation rhythm and see exactly what upgrading unlocks. The benefit list on pricing ("Remove image blur") then matches something they've already wanted. Suggestion chips keep conversations flowing — engagement is the retention engine, and message volume is what makes the subscription feel worth it.

Roleplay games: turning chat into a hook

A flat chat box is easy to leave. So the product wraps conversations in scripted roleplay scenarios — "The Elevator Test," "Beyond the Contract," "Fake Boyfriend" — each presented as a cinematic movie poster with a single Play Game button. It's the same gamification logic as an app's daily-quest screen: give people a reason to open the app that isn't "start from a blank page."

nemora.ai — Roleplay Games
Roleplay game cards styled as cinematic movie posters with Play Game buttons
Live on nemora.ai — scenario cards designed as film posters. Each is a pre-set story that drops the user straight into an engaging conversation.

Why it works

Poster framing sets tone and expectation before the first message, so the chat opens with momentum instead of a blank prompt. It also merchandises the catalog — new scenarios become new "content drops" that pull users back.

The wider retention system

Roleplay games sit alongside an achievements layer and seasonal takeovers (Valentine's, Easter) I also designed — the same playbook mobile games use to convert idle installs into daily habits.

Character creator: complexity behind presets

The creator is the product's emotional hook — and a form with a dozen dimensions. The design keeps it feeling like play, not configuration: visual preset grids for the big identity choices, compact dropdowns for the rest, one sticky "Create" action at the end.

nemora.ai/create
Character creator: visual ethnicity presets and collapsed attribute dropdowns
Visual presets where the choice is emotional, dropdowns where it's parametric — the same progressive-disclosure logic as any good configurator.

Discover: merchandising a content catalog

The home screen works like an e-commerce storefront: a rotating hero campaign, a character grid with personality tags and hover-to-chat, an in-grid "Create your own" card that turns browsing into creating, and an SEO-oriented FAQ block — because in this market, organic search is a major acquisition channel.

nemora.ai
Nemora Discover page: hero campaign banner and character card grid
Discover — campaign hero, character cards, and the create-your-own card embedded in the grid.

Mobile-first: the primary surface

Most of Nemora's traffic is on phones, so mobile isn't a shrink of the desktop — it's the reference design. A persistent bottom tab bar (Discover · Chat · Create AI · Studio · Roleplays) keeps the whole product a thumb-reach away, and the hero, countdown and roleplay cards are all tuned for a single column and one hand.

Mobile home hero — nav, gender tabs, countdown and campaign bannerHome — nav & campaign
Full mobile home with bottom tab barDiscover + bottom nav
Roleplay games on mobileRoleplay games

Captured from the live mobile site. Prefer the real thing? The phone frame up top loads nemora.ai live.

The design system — rebuilt in code

Shipped products need boring rigor: documented brand tokens, opacity ramps, button hierarchies, inputs and tag chips — so a fast-moving team ships consistent UI without me in the room. Below isn't a screenshot; it's Nemora's token set and core components rebuilt in live code — click a swatch to copy its hex, focus the input, tap a tag:

Color tokens — click to copy
Opacity ramp — Nemora Pink
100908070605040302010
Typography
HeadingSpace Grotesk · Bold · 24
Body — teased, not blocked.Inter · Medium · 15
Caption / metaInter · Regular · 12
Components — live, try them
Buttons — hover & press
Input — focus it
Personality tags — tap to select
CaregiverSageJesterDominantConfidant

Why this matters: a live product ships faster when its system is unambiguous. Every state above matches the documented Figma spec — and I can rebuild it in code myself.

Beyond the product: the growth layer

Half of this job was never inside the app:

That's the part I'd underline in this case study: it's designing a business, continuously, not delivering a file.

What I'd do differently

The tension in this product is real: conversion patterns like countdowns and anchors work, but overusing them erodes the trust the product depends on. If I started again I'd define the "pressure budget" up front — which mechanics we use, where, and what we never do — instead of negotiating it screen by screen. And I'd bring cohort data into design reviews earlier: we designed against campaign metrics long before we saw true retention curves.

Next case study
Generato — AI Image Generation
Get in Touch

Let's create something great together

Have a project in mind? I'd love to hear about it.