MEMORIS GLOBAL · STUDIO DOCUMENT · FOR EXTERNAL USE
Ecosystem Inventory and Valuation
CENSUS 9 SEP 2026 · MEASURED, ESTIMATED AND SCENARIO FIGURES LABELLED
Census date 9 September 2026 · Build period 22 October 2025 to 9 September 2026, 322 days · Team one founder, directing AI systems.
Every number in the census sections (02, 03) was measured on the production servers on 9 September 2026 and can be re-verified with the same command; section 05 was re-measured on 27 September 2026. Module counts (04) are our catalogue of 19 August 2026. Valuation ranges (07) and completion shares (08) are management estimates; section 09 is a scenario, not a forecast. Each section is labelled with its kind.
322
Days
5
Platforms
65
Modules
428,351
Lines of code
1,094
API operations
5
Servers
01Why this studio
MEMORIS GLOBAL is an AI studio: one founder directing AI systems, five platforms in
production, one stack (TypeScript, React, Node.js) and one Stripe account underneath, built in 322 days.
Why now. Loneliness has become a declared global public-health epidemic, and
people increasingly choose honest AI companionship over fake profiles and borrowed
faces. That is TRUENECT. The way we are remembered is changing too: in Japan virtually
everyone is cremated, across the West cremation has passed 60% and ash-scattering keeps
rising, ever more people leave no grave, no plaque, no place to be visited. The
memorial moves online, or it disappears. That is MEMORIS. And one AI that actually does
the work, talks, plans, takes web calls, builds whole websites, is SHELA.
Three of the studio's products sit in categories among the hottest in the world right
now. One of those categories saw a two-billion-dollar acquisition battle in 2026.
Included in this census. Between 17 August and 9 September 2026 Shela gained a music arm,
67 MUSIC: a sentence becomes a finished three-minute track with its own cover, an artist keeps the same
face and the same voice across every release, a studio in the browser edits the result down to the note, and a
video editor cuts the clip that carries it. Measured on the production server on 9 September 2026:
27,448 lines of code across 22 client modules and the music server, 445 samples and loops,
11 music and video tools inside Shela. Its lines are already inside Shela's 102,723 above; it is not a separate platform and it changes none of the numbers
above. It is one more reason the same subscription gets spent.
02Measured numbers
Measured on 9 September 2026 · rows marked 16 Aug: inventory of 16 August 2026
What
How much
Product platforms
5
Public web properties (16 Aug)
14
Distinct pages and screens (16 Aug)
~249
Modules, standalone product surfaces (studio)
65
Lines of working code (studio)
428,351
Source code files
2,159
API operations
1,094
Database tables (16 Aug)
179
Interface components (16 Aug)
758
External integrations (16 Aug)
21
Servers in production
5
Lines of documentation (16 Aug)
~55,000
Days of building
322
Rows marked 16 Aug come from the inventory of 16 August 2026; modules are our catalogue of 19 August 2026; all other rows are the census of 9 September 2026. API by type: 740 tRPC procedures + 354 Express routes = 1,094, counted 9 September 2026.
03The five platforms
Measured on 9 September 2026
01Memorismemoris.io160,261 lines · 18 modules · incl. AI Twin (live video coming soon) · in production
02Shelashela.ai102,723 lines · 11 modules + 82 tools · incl. 67 MUSIC and SheAI News · in production
03Truenecttruenect.com63,640 lines · 10 modules · in production
04MailMind / AIMailingaimailing.app29,375 lines · 7 modules · in production
05AIMarketKingmarketing.inki.agency24,733 lines · 8 modules · in production, sign-up and paid plans open, no paying customer yet
Around the five platforms.Stonar (25,930 lines) is the stone business with 25 years in the trade behind it, run today through the founder's company (registered in December 2023, to be renamed Memoris d.o.o. before closing): the partner channel that carries the QR plaques into real cemeteries and opens the trade to Memoris step by step. Memoris Global (11,323 lines) and INKI (10,366 lines) are the studio's own sites. Together with the five platforms that is the 428,351 lines counted above.
Management catalogue, 19 August 2026 (not counted by command)
Every module below is a standalone product surface. Many of them would be a company of their own elsewhere. Where a module is not live today, its state is marked: beta, coming soon or paused.
Memoris · 13 modules (18 with the 5 AI Twin modules below)
Why this matters: three of the products are not in niche categories but in the categories where the most is happening in the world right now.
Category
Market state 2026
Our products
AI companions
$5.2B (2025) growing to $23.7B (2030), 35.4% CAGR. Character.AI: 233M users, $1B valuation. The category earns $20M+ per month. Beneath the numbers: a declared global loneliness epidemic, the demand is people wanting someone who is honestly there.
Truenect
General AI agents
China blocked Meta's ~$2B acquisition of Manus in April 2026 rather than let the technology leave. A year earlier, in April 2025, Manus was valued at about $500M.
Shela
Digital legacy
$12.9B (2024) growing to $30.8B by 2030, 15.6% a year (Grand View Research; other research houses put it higher, we quote the lower bound). Memories.com raised $30M+. 948 death-tech companies. The driver: cremation and ash-scattering keep rising. Japan is at 99.99% cremation, the West past 60%, so ever fewer people have a physical place of remembrance.
Memoris
07Valuation
Management estimate, not a market price
Method A: rebuild cost
If someone commissioned this from an agency: 13 to 19 developer-years.
EUR 1.0 to 2.5 million The lower bound at Slovenian rates (~EUR 45/h), the upper at Western European rates (~EUR 75/h).
Method B: valuation by segment and market weight
Segment
Estimate
Digital legacy (Memoris)
EUR 120k to 350k
AI companions (Truenect)
EUR 60k to 200k
General AI agent (Shela)
EUR 60k to 200k
Marketing OS (AIMarketKing)
EUR 40k to 120k
Vertical SaaS (Stonar)
EUR 40k to 100k
Digital ID + AI Twin
EUR 30k to 100k
Data and distribution
EUR 30k to 90k
Domains, content, brand
EUR 25k to 70k
Other
EUR 10k to 30k
TOTAL
EUR 415k to 1.26M
Management estimate: asset value of the software alone, orderly sale, EUR 415,000 to 1,260,000
What this means against invested capital
Invested: EUR 120,000 (management estimate) plus 322 days of work. At that asset value, it is between three and ten times the amount invested.
An important clarification that must be said plainly: this value is an estimate, and it is unrealized. The asset exists, the liquidity does not. Nobody will pay it out until something is running.
Asset value is not the price of the company. Both methods value the software as it stands today, with no revenue behind it: a floor, not a forecast. The cap in section 10 prices the company going forward, as every pre-seed round does.
08Where we stand
Management estimate · user counts: audit of 2 August 2026 and Stripe on 27 September 2026
Management estimate: an approximate reading of where the whole stands, derived from the measured facts of the estate. It moves every day, in both directions.
Phase
Share
Design and architecture
~100%
Product build
97%
Infrastructure and operations
92%
Content and SEO foundations
80%
Distribution and reach
20%
User acquisition
8%
What is being done right now. The build is finished in the main. The current work is the last pass before launch: security protocols, penetration testing and hardening across all five servers, and the polish that comes with it. Nothing on that list is a new feature. It is the work you do before you invite people in.
No new product is being built, though a builder is never finished. What remains before launch is an age gate, a switch of the video provider and real test purchases on every checkout; the larger work is bringing these worlds to people, so people start seeing us.
Current state: 57 real registered users (audit of 2 August 2026), software payments to date USD 110 (Shela) and EUR 73.97 (Memoris credits), December 2025 to May 2026; 1 active subscription of EUR 11 a month since 20 August 2026 (Stripe). This is not the position of "the product failed". It is the position of "the product has never been introduced". Between those two states lies all the difference.
09What happens when the automation switches on
Scenario, not a forecast
The system was not built to be driven by hand. It was built to switch on in stages, and every stage removes one loop from the founder. No stage needs a new product; what remains is an age gate, a switch of the video provider and real test purchases on every checkout, and then credentials, accounts or a first partner.
Stage
What switches on
Who works after
Status
0
Manual work: every post and contact goes through the founder
a human
behind us
1
Content engine: twelve agents plan, write and design, a linter verifies every claim, a human approves
engine + 20 min/day
built
2
Outbound reach: sequences to 198 selected business addresses (139 funeral homes, 59 stonemasons)
engine
built
3
Syndication: the same approved content on five to seven channels at once
engine
adapters built
4
Partner loop: every funeral becomes a distribution event with roughly forty people
partners
awaiting first partner
5
User loops: a user creates a memorial or a site, a creator brings their audience
users
built
Every stage moves the same boundary: the marginal cost of the next user approaches zero, and the founder's hours stop being the bottleneck. This is the difference between a company that grows by hiring and a company that grows by switching on. The second one was built deliberately.
The first 100 paying users
Scenario, not a forecast. Sections 02, 03 and 05 are measured, 04 is our catalogue, 07 and 08 are management estimates; this section is a projection with stated assumptions. Subscriptions across the platforms run EUR 11 to 799 a month (SHELA priced in USD); the model uses a distinctly conservative blended revenue per paying user.
Blended revenue per paying user
Monthly
Yearly
EUR 12 (cautious)
EUR 1,200
EUR 14,400
EUR 20 (base)
EUR 2,000
EUR 24,000
EUR 35 (favourable)
EUR 3,500
EUR 42,000
The system pays for itself. Servers and AI usage at this scale cost a few hundred euros per month; one hundred paying users cover that several times over.
The funnel is proven. Conversion, acquisition cost and churn become measured numbers instead of guesses, and turn into negotiating instruments.
The valuation changes category. From an asset sale to a company with recurring revenue, the only category buyers and investors take seriously.
One hundred paying users is not a goal because of the money. It is proof that the loop holds.
The first 1,000 paying users
Blended revenue per paying user
Monthly
Yearly
EUR 12 (cautious)
EUR 12,000
EUR 144,000
EUR 20 (base)
EUR 20,000
EUR 240,000
EUR 35 (favourable)
EUR 35,000
EUR 420,000
Market value catches up with build value. At three to five times annual revenue, that is EUR 720,000 (base case, 3x) to EUR 2,100,000 (favourable case, 5x), overlapping the rebuild-cost range of Section 7 (EUR 1.0 to 2.5 million). The gap between what is built and what the market recognizes closes here.
A team becomes possible, not necessary. Two to three hires financed from operations, for acceleration rather than survival.
The negotiating position reverses. The choice becomes capital, acquisition or independent growth, and all three stay open.
Today there is one active paying subscription (EUR 11 a month); all earlier software payments add up to USD 110 and EUR 73.97. The distance to one hundred is not technical but distributional: it requires switching on stages 1 to 4, which are already built. Each is a matter of credentials, accounts and a first partner, not of development.
The stone doorway, why the switches flip in this order
Content engine first. Not because it is the flashiest switch, but because every other
switch sends people somewhere, and an empty channel kills trust faster than no channel.
Content fills the rooms; then outreach and partners bring the guests.
The partner loop comes straight from the stone trade the studio grew out of. Funeral homes and
stonemasons get a QR plaque on the table. The family scans it, creates the memorial on the
spot, and orders the same QR engraved into the real gravestone. A funeral gathers around
forty people, so every stone in nature becomes a doorway: scan leads to a memorial, the
memorial leads to Memoris, Memoris leads to SHELA, and SHELA is the engine that builds
sites and runs life.
One QR plaque turns a cemetery into a distribution network no advertising budget can buy. Stone has been the workshop's distribution channel for 25 years; it just was not called that.
Building was never the bottleneck. Introduction was. The next 90 days fix the
actual constraint, not the comfortable one.
10The raise
Terms of the round (post-money SAFE)
EUR 300,000 pre-seed post-money SAFE · valuation cap EUR 2,000,000 · 15 percent at conversion, before later rounds. One number, one instrument.
Why that price. On our own count, 13 to 19 developer-years at EUR 45 to 75 per hour, rebuilding this ecosystem would cost EUR 1.0 to 2.5 million (management estimate). The EUR 2,000,000 cap sits inside that range and under its upper bound. The asset values in section 7 describe the software alone, with no revenue behind it; the cap prices the company.
The path of the rounds. Management plan, not a commitment: EUR 300,000 pre-seed now, a seed of around EUR 4 million after it, and a growth round about a year later at roughly ten times that level. The first two sit close together on purpose: no new product is being built, so the pre-seed does not pay for construction, it pays for the launch, and the seed follows from measured traction instead of from a plan.
Where the EUR 300,000 goes. Four things, in this order. First hires, so the founder stops being the only pair of hands. Security and the final polish, the work now under way. Distribution: switching on the systems that are already built and checked, the content engine, outbound reach, syndication, creator and user loops, the QR partner loop with funeral homes and stonemasons, and small paid experiments to find the channels that convert. Runway for the founder and operations, plus legal and a reserve. Infrastructure stays a few hundred euros a month:
the capital goes to distribution, not servers.
One milestone. The first 100 paying users. From that point conversion, CAC and churn are measured numbers instead of guesses, and the seed happens from strength.
11The short version
Five platforms, built in 322 days by one founder directing AI, running live on five
servers. Nothing here is a mockup.
Three of them sit in the world's hottest categories: AI companions, carried by a global
loneliness epidemic; general AI agents, the category China refused to let Meta buy out;
and digital legacy, because ever fewer people leave a grave to visit.
No new product is being built, though a builder is never finished. The larger work that remains is bringing these worlds to people: EUR 300,000
pre-seed, post-money SAFE at a EUR 2,000,000 cap, and one milestone, the first 100 paying users.
What is genuinely good about this story: it does not need to lie to be good.
