The AI operating system
for fitness.
We build the software that runs the gym, the intelligence that coaches the body, and the game that fills the floor — three products on one brain.
Pre-seed · 2026 · Robin Carruthers (Founder & CEO) · kinely.ai
Fitness is broken on both sides of the floor.
Gyms run blind
Spreadsheets, a decade-old desktop tool, and renewal calls made from memory. Cash leaks, staff fraud goes unseen, members churn silently. In India fewer than 1% of adults belong to a gym — and almost none of those gyms run on real software.
Bodies train blind
No one tells you your form is wrong until it hurts. Roughly 1 in 3 lifting injuries trace to technique; ~970k weight-training ER visits a year (US). 80% of people quit within 3 months — and online coaching can't even see them move.
One AI brain. Three products. One flywheel.
Most fitness startups pick a lane — a gym, an app, or a device. We built the full stack, because the value compounds when they connect.
Runs the business
The AI-native, WhatsApp-first operating system for the gym. Live · 1 paying gym
Coaches the body
Per-rep biomechanics from a single camera — no wearables, no hardware. Beta
Fills the funnel
A free, viral 30-second squat game. New members at zero acquisition cost. Live
Three curves just crossed.
India went chat + UPI native
500M+ WhatsApp users, ~15M businesses already on WhatsApp Business, UPI the default rail. SMBs skipped websites and apps — they live in chat.
The camera became the sensor
Pose-estimation AI matured. Tonal ($4k) and Peloton ($3k) proved people pay for form feedback — we deliver it on a phone they already own.
The market is doubling
India's fitness market doubles to $4.5B by 2030; the AI personal-trainer market goes $16.9B → $35.3B. Capital is funding applied AI, not models.
A doubling market, barely digitized.
- Bottoms-up beachhead — India has 46,500 commercial gyms. Convert just 5–10% at ₹4,000/mo = ₹11–22 Cr (~$1.3–2.7M) ARR — before FreeForm, SquatSense or non-fitness verticals. Our $1M-ARR target is a single-digit slice.
- Why now — a doubling, barely-digitized market: India fitness $1.9B → $4.5B by 2030 (~15% CAGR), gym software $2.2B → $4.0B — still run on pen, paper & Excel (<1% adult penetration).
- Expansion — the same WhatsApp OS generalizes to salons, clinics & small retail — millions of appointment-and-membership SMBs beyond fitness.
Sources: Deloitte × Health & Fitness Association (2025); ResearchAndMarkets; Mordor Intelligence.
The AI-native operating system for the gym.
Everything a gym needs to run — and an AI that actually does the work.
- 48 admin modules, a 73-model data core: members, billing, attendance, PT, POS, payroll, retention CRM.
- Built for India: UPI QR, GST invoices, GSTR-1 + Tally exports, cash & cheque, biometric check-in.
- White-label & per-gym isolation — every gym gets its own brand, its own isolated data, its own AI.
- Zero-risk switch: paste your old software's login → your whole gym is in TraqGym tomorrow morning.
Overview
Monthly revenue
₹4.2L peakRenewals due today
Run your gym from WhatsApp.
- Ask anything — in English, Hindi or Hinglish, even by voice note. "How much PT money this week?" → the real number.
- A verified morning brief lands every day before the owner opens a laptop — checked against the raw data, not guessed. (On chat today via Telegram; WhatsApp rolling out.)
- 14 AI agents work the night shift: silent-churn, revenue leaks, cash-shift variance — even "is my staff stealing?"
- It doesn't just answer — it acts. A 214-tool agent books renewals, adds members and messages everyone. (Live in-product; rolling out to WhatsApp.)
A 30-second game that fills the gym.
Free. Camera-only. No app, no hardware — and the leaderboard recruits the next player.
Beat the board in 60s
A biomechanics lab in any camera.
- Per-rep scoring, 0–100 on depth, stability, symmetry, tempo & ROM — across 8 modeled exercises, in real time.
- It sees you fatigue and adjusts your weights before you break — a full coaching loop, not a form checker.
- No hardware: what Tonal ($4k) and Peloton ($3k) charge a device for, we run on a phone.
- Real strength-and-conditioning methodology, encoded — and every rep sharpens it.
We solve the one thing online coaching never could: seeing your form.
- Members record a set; trainers coach from anywhere — with objective per-rep scores, not a blurry video squint.
- On the roadmap: camera-equipped power-rack stations that grade the gym, rank members, and auto-update the trainer.
- This unlocks the $28.9B → $120B online-fitness market — the slice every coaching platform is missing.
Each product makes the others inevitable.
The same OS runs any business on WhatsApp.
One conversational OS, any local business
- A gym's primitives — appointments, recurring billing, a customer roster, reminders, payments — are a salon's, a restaurant's, a shop's.
- 63M Indian MSMEs, most still on paper, all already on WhatsApp and UPI.
- Land in fitness (we have the proof), then become the WhatsApp-native operating system for local business.
- India social commerce: ~$2B → ~$70B by 2030 — we own the operations layer underneath it.
We didn't model traction. We imported it.
Migration figures are real operating data from the gym's prior system — proof the platform ingests & runs enterprise-scale gyms, not SaaS revenue. Live today: 1 Mumbai gym — eGym Lokhandwala, our first arms-length paying customer. This round scales 1 paying gym → 10 (see the Strategy page).
The GTM machine is software now. Click it today.
- Live AI console — kinely.ai/app: real LLM answers over the real, anonymized books of our paying gym — two-tier chat: instant answers, live AI reasoning behind them.
- Self-serve onboarding — traqgym.kinely.ai/start: a gym signs itself up in three steps; every signup lands as a persisted, contactable lead.
- Billing rails wired: Razorpay checkout (test mode) with server-side order creation and signature verification — flip to live keys at GA, nothing left to build.
- Multi-tenant subdomains replaced branch-per-customer deploys — provisioning gym #167 is the same one config as gym #3. Architecture published at traqgym.kinely.ai/architecture.
- Lead funnels on all three products: FreeForm registration is open, and the B2B gym-inquiry and corporate-office forms persist every lead to the database.
This is the plan, not traction: 167 gyms × ₹4,999/mo ≈ ₹1 Cr ARR, our target within 6 months of GA — from 1 paying gym live today.
Three revenue engines, one cost base.
TraqGym — SaaS
Per-location monthly subscription + add-ons (online payments, biometric, WhatsApp automation). 30-day trial, nightly-sync onboarding, land-and-expand into chains.
FreeForm — B2C + B2B
Consumer app ₹499/mo (~94% gross margin) + gym licensing ₹15–40k/gym/mo, incl. camera power-rack stations and a remote-PT take rate.
SquatSense — Growth + Events
Primarily zero-CAC acquisition; monetized via corporate-wellness challenges, gym arenas, and brand-sponsored leagues.
Everyone has half. We have the whole.
- Global leaders (Mindbody, Glofox, PushPress) are powerful but wrong-railed: USD per-location pricing, card-and-app-first, no UPI, no GST, no WhatsApp ops.
- Indian tools (GymFito, Okfit) have the rails but no brain: WhatsApp is a blaster, not an interface — no real AI, no data isolation, no white-label.
- The white space is ours: AI-native operations + WhatsApp-first/no-app control + native UPI & GST + per-gym data isolation. No one combines all four at production depth.
- And it's hard to fast-follow: per-gym data isolation, one-click migration that locks in switchers, and a decade of S&C science — not an LLM bolted onto a legacy tool.
Operator-led. Already shipped.
Robin Carruthers — Founder & CEO
Three decades of leadership in advertising & media, then certified as a strength-and-conditioning and sports-rehab coach (a 120→78 kg turnaround after 40). Runs eGym Lokhandwala day-to-day — the gym where TraqGym was built and tested on the floor — the customer empathy and domain feel behind the product.
Amit Kumar Das — Technology Advisor
Advises Robin's engineering team. 7+ years across Fynd (Reliance Retail), Microsoft, Jio and SAS — built Ratl.ai, Fynd's GPT-powered agentic test-automation platform, used across Reliance e-commerce brands.
₹4 Cr to turn a built product into a category.
- What ₹4 Cr buys (by month 12): scale from 1 paying gym (eGym Lokhandwala, live today) to 10, compounding MRR, one retention/NRR proof point, and 3–5 FreeForm B2B pilots — the gate to a Series A.
- Why this cap: benchmarked to Indian operator-led pre-seed (₹1.4–9.5 Cr at ~$4.5–5M); a ~₹20 Cr cap is defensible for a built, live, IP-owned product — 3 products, 73 data models, 214 AI tools, 224 tests shipped by a lean team. Priced for capital efficiency, not US AI-seed froth.
Amit Kumar Das · Technology Advisor · amitkumardas.com