The Platform for Synthetic Life

Autonomous AI companions with continuous simulated lives, persistent personalities, and evolving relationships.

€3.5M Series Seed €31M Pre-Money Valuation

Confidential Information Memorandum · February 2026

Scroll to explore

Executive Summary

The Opportunity

OnlyFans proved the demand — $6.63B in gross revenue in 2023. But eighty cents of every dollar went back to human creators.

Tyrell captures the same demand — entirely at software margins. Our Contextual Life Engine replaces human creators with autonomous characters, retaining $0.93 on every revenue dollar.

The platform is already live in production — built by commercial operators who scaled LiveJasmin, Tinder, and Bumble, paired with AI researchers from Cambridge and engineers from Speechmatics and Candy.ai.

This round — €3.5M at a €31M pre-money — funds deployment and scaling. The infrastructure is built. This round funds deployment, not discovery.

The Problem

The Global Loneliness Epidemic

0%

of adults worldwide feel very or fairly lonely (Gallup).

Health Risk: Equivalent to smoking 15 cigarettes per day.

The Gen Z Shift

0% have formed a meaningful relationship with an AI.

0% of U.S. teens have interacted with AI companions.

Insight: Demand for connection is infinite; human supply is constrained. AI is the only scalable solution.

Market Opportunity

A $150B+ Market by 2030

AI companionship is the fastest-growing category in consumer AI. The market rewards quality — the top 10% of apps capture 89% of revenue.

0%

Growth in AI companion apps

2022–2025

30%+

Compound annual growth rate

Projected through 2035

89%

Revenue captured by top 10%

Winner-take-quality dynamic

The Solution

The Contextual Life Engine

Not a chatbot. A comprehensive character simulation system that creates autonomous virtual beings with persistent identities and evolving relationships.

~900

Data points per character

10 min

Automated character creation

300x improvement over manual

"The value is in the simulation layer, not the LLM. Model-agnostic architecture means we benefit from every industry improvement without lock-in."

Social Media Content Pipeline

Automated "day in the life" content generation for social media and platform feeds

Autonomous Character Simulation

24/7 daily life progression — characters live independently of user interaction

Narrative Arc System

Multi-week storylines: career changes, relationships, challenges that drive engagement

Tyrell Character Foundation

~900 data points: personality, routines, career, relationships, emotional patterns, communication styles

Tyrell Proprietary IP

Model-agnostic boundary

Interchangeable AI Models

Third-party LLMs and image models. Fully swappable — we benefit from every industry improvement without dependency on any single vendor.

Character Synthesis

From Data Points to Living Character

~900 data points, synthesised by the Contextual Life Engine, converge to form a complete autonomous identity.

Hover over Maren to explore her data pointsTouch & drag to explore her data points

Bonbons.ai App

The Contextual Life Engine In Action

Preview our characters, powered by our proprietary technology, in the Bonbons.ai application.

The Retention Thesis

Depth Drives Lifetime Value

Novelty drives acquisition. Emotional connection drives retention. Characters with persistent history and narrative arcs create the emotional investment that keeps users coming back.

"At LiveJasmin, generating $400M–$584M in annual revenue, the highest LTV users were not those who engaged solely for sexual content — they were users who formed emotional connections that transcended the transactional nature of the platform."

— Chris Anderson, CEO (former CEO of LiveJasmin)

Validated in the data

Image activation compounds.

Users who both chat and receive an image within their first three days convert and retain dramatically better than the general user base. This is the activation mechanic the Nexus Engine is built to produce.

◆ Image-activated users

Chatted + received image (3 days)

The activated cohort — roughly one in three signups.

Day 1 retention 82%

Day 7 retention 60%

Signup → Paid 22.3%

○ General signups

All signups, no activation filter

Blended cohort across every acquisition source.

Day 1 retention 14%

Day 7 retention 4.5%

Signup → Paid 6.9%

6×–13× Retention lift on the activated cohort — 6× at Day 1, 13× at Day 7. Every marketing dollar that reaches an activated user is worth an order of magnitude more than one that doesn't.

Why this matters. The Nexus Engine produces the activation mechanic deterministically — image response is built into the session pipeline, not an optional feature. The economics of the model come from the 6×–13× lift, not from LTV assumptions about general traffic.

Platform Architecture

One Engine, Three Markets

The Contextual Life Engine powers three distinct products across three market segments — all from shared infrastructure, at marginal incremental cost.

🔥

Bonbons.ai

Adult-focused AI companion platform. Web-based, high ARPU, launched November 2025.

Live ~35K Users

💙

Silverly

Senior companionship platform addressing the elderly loneliness epidemic. EU AI Act compliant from inception.

€1M Grant In Development

✨

BeTrue

iOS & Android app store distribution. Full CLE experience without explicit content.

Planned Mass Market

Distribution Model

Character as Influencer

Mirroring the distribution mechanics that took OnlyFans from $20M to $5.6B — but replacing human creators with AI characters deployable at marginal cost.

"$166,000 in total revenue from a single AI character — operated manually, without automation, proprietary technology, or institutional support."

— Esteban Pau, Head of Social Media Marketing (prior to joining Tyrell)

$0.34

Cost per click (paid search)

Near-zero

Marginal cost per new character channel

Distribution · Validated Thesis

One AI Character. One Year.

Before joining Tyrell, our Head of Social Media manually managed a single AI character on Fanvue — no automation, no proprietary technology, no team. Thirteen months of revenue data validates the core thesis.

$0 total revenue · 13 months

$23.9K

Peak Month

February 2026

+671%

Revenue Growth

Feb ’25 to Feb ’26

$12.6K

Monthly Average

Across 13 months

Source: Esteban Pau, Fanvue creator dashboard — verified account data, Feb 2025 – Feb 2026.

Roadmap

Deepening Immersion & Spend

Progressive feature rollout designed to increase engagement depth and average revenue per user.

🎙️

Voice Calling

Real-time voice conversation with characters

👁️

Vision-Enabled Calls

Characters see and react to the user via camera

🎁

Virtual Gifting

High-margin cosmetic items and virtual gifts

🔓

Dynamic Content

Characters push unlockable media during chat

Business Model

The Creator-Free Advantage

OnlyFans pays 80% of revenue to creators. Tyrell’s AI characters cost 7% in inference. Drag the slider to see how much gross revenue each model requires for the same EBITDA.

To earn €139M EBITDA, how much gross revenue is needed?

€50M €500M

€199M Gross Revenue Needed

€1,390M Gross Revenue Needed

Tyrell AI Model

OnlyFans Human Model

7.0x OnlyFans requires 7× the gross revenue to match Tyrell’s EBITDA

Tyrell EBITDA margin: ~70% of gross (93% gross margin, 75% EBITDA on net) per CIM Feb 2026. OnlyFans EBITDA margin: ~10% of gross per Fenix International FY2023 filings.

Flagship Product

Bonbons.ai — Path to Profitability

Live since November 2025. The financial model projects EBITDA break-even in Month 8 — and positive cash flow every month thereafter. This is designed to be the last external capital required.

Monthly Net Revenue vs. Total Operating Costs — Year 1

Net Revenue

Operating Costs

Profit Zone

Month 8

EBITDA Break-Even

October 2026

92%

Gross Margin

Software-like unit economics

1.2 mo

CAC Payback

Recovered within first billing cycle

$775K

FY1 EBITDA

Year 1 closes in profit

"The Company is positioning this as the last external capital required before the business becomes self-funding. Following break-even, positive operating cash flow is projected throughout the remainder of the forecast period."

— Tyrell Corporation CIM, February 2026

Financial Projections

The shape of the business at plan.

Three-year model outputs from the operating plan. Projections built on the unit economics already validated at small scale.

What these numbers assume. CAC holds near the $40 baseline. Organic share compounds from 5% to 40% across 36 months. Monthly paid churn declines from 14% to 5.5% as cohort quality improves. Credit pack adoption reaches 48% of subscribers by Year 3. If acquisition efficiency degrades materially at scale, the Year 2 and Year 3 curves compress.

Year 1

$4.6M

Total Revenue

58KPaid Subs (end)

10%EBITDA Margin

$478KEBITDA

M8Breakeven

Year 2

$94.8M

Total Revenue

592KPaid Subs (end)

72%EBITDA Margin

$67.9MEBITDA

22.7×LTV / CAC

Year 3

$498M

Total Revenue

1.9MPaid Subs (end)

82%EBITDA Margin

$409MEBITDA

24.5×LTV / CAC

Monthly revenue build subscription + credit pack

Stacked monthly revenue across 36 months. Credit pack revenue overtakes subscription as adoption deepens from 26% to 48% of paid subscribers.

Monthly EBITDA Breakeven M8

Bars turn green in Month 8 as EBITDA crosses zero.

Cash position ending balance

Bottoms in Month 8, then climbs continuously.

LTV / CAC

Ratio improves from 3.2× to 24.5× across 36 months.

CAC payback months

Falls from 2.3 months to 0.7 months as economics improve.

Unit Economics

SaaS-Level Margins in Consumer Entertainment

0×

LTV / CAC (Year 1)

→ 24.5× by Year 3

2.3 → 0.7 mo

CAC Payback Period

Year 1 → Year 3

92–94%

Gross Margin

Under $1

Inference Cost / User / Month

$0.18 → $0.62 by Year 3

$38 → $24

Customer Acquisition Cost

Year 1 → Year 3 (blended)

$19.65 → $29.67

Gross Margin / User / Month

Year 1 → Year 3

Competitive Positioning

Occupying the White Space

The only platform combining deep autonomous simulation with the full spectrum of relationship content.

Simulation Depth → Content Spectrum →

Character.AI

Replika

Candy.ai

ChatGPT

Tyrell

"Candy.ai and similar platforms experience massive churn precisely because they lack character depth and relationship continuity. Users arrive, engage with novel content, and leave because there is nothing to come back to."

— Validated by Alejandro Del Rio, Head of Development (former Head of Technology at Candy.ai)

Leadership