Skip to content
View ZWISERFIT's full-sized avatar

Block or report ZWISERFIT

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ZWISERFIT/README.md

⚡ ZWISERFIT

The AI Operating System for Physical Businesses. One brain (Momo). Two services (Saros + Melody). One data layer (KinTwin). One protocol (Zeus).

We're not building an AI company. We're building an AI OS that runs companies.

每一个Agent都是一个操作系统。不是"用了AI的公司"——是每一个细胞都是AI的公司。

📖 Read this document in: [English (current)] | 中文

YC Summer 2026 AI Operating System AI-Native Service AI Personalized Medicine Software for Agents Company Brain Offline Physical Health Infrastructure

1 node · 7 years · 118 active members · 9 AI Agents · 24/7 autonomous · 5 YC RFS directions


30-Second Elevator Pitch

ZWISERFIT is an AI OS for companies, built from 9 vertical Company Brain Agents. The front-end delivers two AI-native services: Saros store digital co-founder solves Planet Fitness' labor cost problem; Melody personal metabolic coach manages energy, glycolipid, and hormonal health in three dimensions. The bottom layer, KinTwin, captures hardware-verified continuous behavior streams — more objective and unforgeable than Nourish's self-reported data. All encrypted assets flow through Zeus Protocol to insurance institutions for dynamic risk pricing. Five YC RFS directions covered in one company.

ZWISERFIT 是一套由 9 大垂直领域 Company Brain 组成的企业级 AI 操作系统。前端两套 AI 原生服务:Saros 门店数字合伙人,帮 Planet Fitness 这类连锁健身房重构人力成本;Melody 个人代谢管理教练,帮会员实现能量+糖脂+激素三维代谢管理。底层 KinTwin 连续行为流基础设施对标 Nourish,但数据来自线下门店——更连续、更客观、不可伪造。全部加密资产通过 Zeus 协议栈对接保险机构,实现动态风险定价。一次覆盖 YC 五条赛道。


💰 For Investors: Deck → · Live Data ↓ · Core Moats ↓ 🏪 For Gym Owners: Saros Open → · Deployment docs 🤖 For AI / Protocol Developers: KinTwin SDK → · Zeus Specs → 📜 Founder Full Story (中文): → FOUNDER_STORY.md


What ZWISERFIT Is

Four $B+ companies. Four separate bottlenecks. One OS.

  • Planet Fitness (2,795 locations): 40%+ of revenue goes to labor — eating 2/3 of profit.
  • Life Time ($900M CAPEX): Right direction, wrong leverage — 500 coaches don't scale.
  • Nourish ($215M raised): Proved insurers want health data — but data is self-reported. Unverifiable.
  • WHOOP / Oura: Data belongs to the platform, not the user.

None of the four industry leaders owns the complete closed-loop pipeline: offline behavior collection → personalized metabolic intervention → cryptographic data asset confirmation → automated insurance risk pricing monetization. ZWISERFIT does. With one AI brain.

Most vertical fitness/health startups only build single-layer tools. ZWISERFIT is the first to integrate all five layers as one platform.

ZWISERFIT = YC RFS #15: AI Operating System for Companies — the only category that can carry all five layers as a single integrated platform.

Layer Product What It Does Unlocks YC RFS
Application Saros + Melody Store co-founder (B2B) + Metabolic coach (B2C) 2.5× profit per store #2
Data KinTwin Hardware-verified continuous behavior streams Insurance-worthy data #3
Kernel 9-Agent Cluster 24×7 autonomous company operations Zero human ops overhead #4
Protocol Zeus Protocol Cross-agent data → automated insurance pricing Agent-to-agent revenue #12

One Brain, Two Faces

Momo is one of the 9 Agent OSes — specifically, the Store O/S. She is the "brain" of Saros and Melody, not the commander of all 9 Agents. (Shuyu is the Strategy O/S that orchestrates the full kernel.) Two external faces for different users.

Momo(Store O/S · One of 9 Agent OSes inside the Company Brain kernel)
├── Saros → B2B:Store Digital Co-Founder(replaces store manager labor cost)
└── Melody → B2C:Personal Metabolic Coach(replaces nutrition coach headcount)

Momo is the vertical store-operation Company Brain within the 9-Agent kernel (YC #4). It is the unified cognitive engine powering both B2B Saros and B2C Melody, with separated execution interfaces for different user groups.

  • For members: They only know "Momo." "Momo planned my workout" (Melody). "Momo checked me in" (Saros). They don't need to distinguish.
  • For store owners: Momo runs the store as Saros — the B2B external interface built upon Momo's store operation kernel.
  • For investors: Two independently-valuable business lines — B2B SaaS + B2C health management — running on the same kernel.

One brain, two faces, one brand. Members know "Momo" as their coach. Store owners know "Momo" as their store manager. Investors see two revenue lines on one kernel. This is not complexity. This is elegance.


The Competitive Stack: Four Companies, One Stack

Company What They're Stuck On Our Move
Planet Fitness Labor eats 40% of revenue → kills profit Saros replaces store manager — because 40% labor is the largest fixed cost; removing it unlocks 2.5× profit
Life Time Right direction, wrong leverage — 500 coaches don't scale Melody — zero marginal cost per member, because AI coaching cost is fixed regardless of user count
Nourish $215M validated the market — but data is self-reported KinTwin — hardware-verified continuous streams, because offline behavior cannot be faked
WHOOP / Oura Data belongs to the platform, not the user Zeus — user-owned data → insurance pricing, because ownership unlocks monetization
⬆️ All four combined None have the full closed loop: offline behavior → metabolic intervention → insurance monetization ZWISERFIT has all four layers

Saros: Store Digital Co-Founder

Saros is the B2B external interface built upon Momo's store operation kernel.

Momo(Store O/S · Founder's cognitive encoding)
    +
SaaS Stack(POS / CRM / Scheduling / Membership / Inventory)
    =
Saros — 门店数字合伙人

Saros delivers end-to-end autonomous store operation — not just standalone SaaS tools for manual human operation. The AI runs the store; humans deliver warmth.

YC Direction #2: AI-Native Service — 2.5× profit per store by replacing the largest fixed cost: store manager labor.


Melody: 3-Layer Metabolic Coach

GLP-1 is opening the floodgates. Millions are on weight-loss medications. But here's the paradox: without exercise, users lose muscle — not just fat. Muscle loss → lower basal metabolism → faster rebound after stopping → worse than before. Insurers are starting to pay for interventions that prevent this. The question is: whose data do they trust?

Melody manages metabolism across three layers, powered by the same Momo kernel:

Layer What Accessible To
① Energy Metabolism Caloric balance · Body fat % · Muscle mass · BMR ✅ Everyone
② Glycolipid Metabolism Blood glucose · Lipid profile · Visceral fat · Insulin sensitivity ✅ Everyone
③ Hormonal Metabolism Menstrual cycle hormones · Cortisol rhythm · Thyroid function 🔥 Women only

"Women's hormonal lifecycle is systematically ignored as a core variable in cardiovascular disease risk modeling."

American Heart Association, Circulation 2024 Official Scientific Statement

There is a 30% structural gap in health data infrastructure for women. Questionnaires cannot fill it. Wearables alone cannot fill it. It requires offline continuous behavior streams + metabolic biomarkers + hormonal cycle tracking — the exact three-dimensional pipeline that only KinTwin captures.

No competitor can replicate the offline hormonal cycle tracking pipeline.

YC Direction #2: AI-Native Service — zero marginal cost metabolic coaching, powered by the same Momo kernel.


KinTwin: Unforgeable Behavior Data

Facial-recognition Gate ──→ Cameras ──→ Body Scanner
         │                      │              │
         └──────────────────────┴──────────────┘
                        │
              Hardware Root of Trust
                   (Physical Layer)
                        │
              Nova Continuous Behavior
                Stream Generation
                        │
              Ethan SHA-256 Notarization
                + Ownership Anchoring
                        │
              KinTwin — Crypto-verified
              Behavior Data Asset Pool
Nourish ZWISERFIT (KinTwin)
Data source Self-reported questionnaires Hardware capture (gate/camera/scanner)
Verifiability Unverifiable (self-report) Unforgeable (biometric identity + hash notarization)
Frequency Static, single-point Continuous, daily — every gym visit = one behavior record
Dimensions Incomplete (only what users choose to report) 3D complete (energy + glycolipid + hormonal)
Sample Subjective self-reported user input Hardware-authenticated objective long-term behavioral records

Nourish validated the demand — but it can never deliver the real answer. KinTwin delivers what insurers actually need: unforgeable continuous behavior streams with cryptographic proof. ✅ Stella Compliance Audit Verified

YC Direction #3: AI-Native Personalized Medicine — Angel round narrative entry point. The most tangible, investable wedge into the full platform story.


Zeus Protocol: Behavior Data → Insurance Pricing

Internal 9-Agent Cluster
    │  Ethan hash ciphertext → unreadable to external agents
    │
    ├── Zeus ── Sole Authorized Oracle Decoder
    │           Actuarial data packets / Valuation interface / Asset credentials
    │
    ▼
External Agents (VC DD · Insurance Underwriting · Pharma Procurement)

GLP-1 is opening the floodgates for insurers to pay for health outcomes. Nourish proved insurers will pay for nutrition interventions. Zeus proves insurers will pay for exercise-verified health outcomes — and exercise can only be verified through offline, continuous behavior streams.

Revenue path:

Stage Model Revenue Type Reference
Now Hardware + Saros/Melody subscriptions Hardware + SaaS ARR Planet Fitness membership model
Mid-term Insurance data products ¥50-250/person/year data licensing Nourish insurance data monetization
Long-term PoPB (Proof-of-Behavior Protocol) licensing + cross-industry data market Protocol fee + transaction commission Visa-style 0.3% per transaction commission model

PoPB (Proof-of-Behavior Protocol) — a cryptographic standard that proves "this person actually exercised" using hardware-verified, SHA-256 anchored behavior streams. Insurers verify physical facts, not user self-reports.

Distinction from Tristan Agent: Tristan maintains hardware and network physical transmission pipelines. Zeus exclusively defines actuarial data format, Agent handshake rules, and asset transaction state machines. Tristan guarantees connectivity; Zeus enables monetization.

YC Direction #12: Software for Agents


9-Agent Kernel: Company Brain

9 independent vertical brain OSes run in parallel. 24×7 self-scheduling, self-auditing, self-iterating.

Founder Suzanne Mok(Cognitive Origin)
    │  Decade of cross-domain frontline wisdom — irreplicable
    │
    ├── Strategy ──── Shuyu    Strategy O/S · Enterprise command
    ├── Store ─────── Momo     Store O/S · Physical operations (founder's family name "莫")
    │
    ├── Data ──────── Nova     Behavior Stream O/S · Continuous data generation
    │                 Ethan    Sovereignty O/S · Hash notarization + DID
    │
    ├── Protocol ──── Zeus     Capital Protocol O/S · External I/O gateway
    │
    ├── Operations ── Tristan  Infra O/S · Hardware + network transmission
    │                 Baron    Content O/S · Brand + narrative
    │                 Luna     Community O/S · Developer ecosystem
    │
    └── Audit ─────── Stella   Compliance O/S · Independent verification

9 OSes. One Kernel. Zero human operational overhead.

Each Agent is a complete, independently-running operating system — with its own scheduler, file system, external protocol stack, and self-iteration loop. The founder's decade of cross-domain frontline wisdom is encoded as the boot sector of every Agent OS.

YC Direction #4: Company Brain


YC RFS Coverage

# YC RFS Direction ZWF Module Role
#15 AI Operating System for Companies ZWISERFIT PRIMARY · Top-level container
#2 AI-Native Service Companies Saros + Melody Acquisition funnel
#3 AI Personalized Medicine KinTwin Revenue engine (angel narrative entry)
#4 Company Brain 9-Agent Cluster Kernel moat
#12 Software for Agents Zeus Protocol Revenue pipeline

#15 is the primary track. It is the only category that can carry all five layers — 9-Agent kernel + Saros app + Melody app + KinTwin data + Zeus protocol — as a single integrated platform. The other four directions map to specific modules inside the OS.


Three Irreplicable Core Moats

  1. Hardware Root of Trust — Self-developed facial-recognition turnstile + cameras + body scanners. Data integrity starts at the physical layer, not the database. No software-only solution can replicate this.
  2. 7 Years of Unforgeable Data — 380,000+ SHA-256 anchored behavior records from Dongguan Wanjiang, China's lowest fitness penetration. You can fork code. You cannot fork 7 years of continuous, cryptographically-verified behavioral data. Time is the only moat competitors can never compress.
  3. Founder-Encoded OS — Each Agent OS boot sector was hand-coded with the founder's decade of cross-domain frontline wisdom — PwC audit × Ping An Life insurance pricing × 7 years gym operations × AI system creation. Not deployed. Not integrated. Born from real-world trial and error. No competitor can replicate this — not because they can't hire smart people, but because no one else has lived the full stack: audit × insurance pricing × gym operations × AI system design, in one person, over one decade.

Live Verified Operational Data ✅ Stella Compliance Audit Verified

Metric Value What This Means
Active monthly members 118 Proven in the hardest market (Dongguan Wanjiang, China's lowest fitness penetration — national average is 3× higher) — if it works here, it works anywhere
Daily check-ins 35–55 7 years of continuous, unforgeable behavioral data
Total verified sessions ~380,000 Each session SHA-256 anchored — insurance-grade audit trail
AI Agent uptime 24/7 Production-ready since March 2026
Data pipeline latency <2s Real-time, not batch — insurance pricing in motion

The Founder: Why This Exists

莫淑瑜 (Suzanne Mok)

In 2017, the founder lost a family member to cancer — the same person who most opposed her entering insurance. She joined insurance to "put a financial firewall under every family." After the loss, she asked: what matters more — money or people? She left insurance. Started a gym in Dongguan Wanjiang — China's lowest fitness penetration. 7 years. One store. Through COVID, through losses, through debt. Not because she couldn't expand. Because she refused to expand before the OS was ready.

In 2025, she completed the full PRD defining every layer of what would become ZWISERFIT — before AI went mainstream. In 2026, she built 9 AI Agents by hand, replacing an estimated 26 weeks and millions in software engineering. She gave her own name to two Agents: Shuyu (Commander) and Momo (Store Manager, family name "莫").

The founder didn't catch up to AI. AI caught up to the blueprint she drew in 2025.

Her four identities — PwC auditor · Ping An Life Insurance 5 years (insurance pricing insider) · 7 years gym operations · AI system creator — are irreplicable. All 9 vertical Agent operating systems are built on the founder's exclusive cross-industry cognition covering offline venue operation, insurance underwriting and risk audit — a cognitive moat that cannot be replicated by competitors relying solely on technical algorithms. This is not a team. This is a cognitive origin.

The founder didn't use AI. She gave birth to it.

Read the full founder story (中文) →


Open Source Strategy: Ecosystem First, Moat Protected

Open the application layer to build ecosystem. Close the kernel and protocol to protect the moat. The moat isn't in code — it's in 7 years of data, hardware trust root, and time.

Layer Strategy License What's Open Moat Protection Logic
🟢 Saros Full Open MIT Momo SaaS stack, deployment, docs Open store entry to scale offline network; proprietary Momo decision logic remains closed
🟢 Melody Full Open MIT Metabolic calculation SDK, coaching engine Lower barrier for member adoption; hormonal cycle proprietary model locked
🟡 KinTwin SDK Open MIT (SDK) / Proprietary (Core) Hardware API specs, data format, integration examples Standard hardware access without exposing Nova behavior stream algorithm
🔴 Zeus Protocol Specs Open CC BY-NC-ND 4.0 Agent communication standards, API docs Prevent competitors copying insurance data trading business logic
9-Agent Kernel Docs Only Proprietary Architecture diagrams, design philosophy Founder's 7-year operational cognitive encoding; irreplicable core brain

Legal Note: Saros, Melody, and KinTwin SDK are MIT-licensed. Zeus Protocol specification adopts CC BY-NC-ND 4.0. The 9-Agent kernel source code remains proprietary and closed.


Repository Structure

New here? Start with /saros-open/ to deploy an AI store manager, or /investor for the full deck.

ZWISERFIT/
├── 🟢 saros-open/          MIT — Store digital co-founder (deploy to any gym)
├── 🟢 melody-open/         MIT — Metabolic coaching engine (integrate into your app)
├── 🟡 kin-twin-sdk/        SDK open — Hardware API + data format specs
├── 🔴 zeus-protocols/      Specs open (CC BY-NC-ND) — Agent communication standards
└── ⚫ 9-agent-kernel/      Docs only — Architecture + design philosophy

What You Can Build On

Directory What You Can Do
🟢 Saros Deploy an AI store manager to any gym. MIT.
🟢 Melody Integrate 3-layer metabolic coaching into your app. MIT.
🟡 KinTwin SDK Connect your hardware to our behavior data pipeline. MIT (SDK).
🔴 Zeus Protocol Build agent-to-agent data交易 applications. Specs open.
⚫ 9-Agent Kernel Architecture docs only. The core brain stays closed.

Target Stakeholders

You Are Start Here
💰 VC / Investor Investor Deck · Live Data ↑ · Core Moats ↑
🏪 Gym Owner / Operator Saros Open · Deployment docs · (2.5× profit case study coming soon)
🤖 AI / Protocol Developer KinTwin SDK · Zeus Specs

Contributing

We welcome developers, venue operators, and researchers — especially to the open-source Saros and Melody layers.

  1. 🐛 Found a bug? → Open an Issue
  2. 💡 Have an idea? → Start a Discussion
  3. 🔧 Ready to code? → Pick a Good First Issue

Links

Into Link
🚀 Product Hunt producthunt.com/posts/zwiserfit
📖 Investor Quick Start → investor
🤖 AI Agent Army (24×7) → agents
🔗 PoPB Protocol Spec → protocol
📜 Founder Full Story (中文) → FOUNDER_STORY.md
🛡️ Stella Audit Signature ✅ VERIFIED

ZWISERFIT · Dongguan, Guangdong, China AI + Human Synergy. Users own their data.

📄 MIT

Popular repositories Loading

  1. ZWISERFIT ZWISERFIT Public

    大健康行为数据基础设施 · AI×Web5 · PoPB v1.0 MIT开源 · 用户拥有自己的健康数据 · 7 years live · 9 AI Agents 24/7

    HTML 1 1

  2. zwiserfit-ai-store-manager zwiserfit-ai-store-manager Public

    Momo — 首个AI门店店长 · 物理AI的应用 · 微信就是门店

    Shell 1 2

  3. PoPB PoPB Public

    PoPB Protocol Specification — Proof of Physical Behavior. The verification protocol for human physical activity. Apache 2.0 + MIT.

    Python 1

  4. .github .github Public

    ZWISERFIT Organization — community health files, contribution guides, and PDL license

  5. investor investor Public

    ZWISERFIT Investor Deck — 7-year traction data, PoPB protocol, token economics, and ILM convergence thesis

  6. data data Public

    7 Years of Verified Physical Behavior Data — 100% capture rate, 118 active members, PoPB-anchored