Members don't churn because of price. They churn because of the cable row.
Liftline is an AI operations platform for gyms — real-time equipment status, smart member queues, NLP-triaged issue reports, and predictive maintenance across three role-based portals. Your workout, uninterrupted.
3
Role-based portals
<100ms
Real-time updates
13
Claude AI features
Pilot
Charter slots open
See it running
Two-minute tour through the live product.
~23 MB · MP4 · sound optional · tap to play
The hidden cost
The reasons members leave are operational, not emotional.
Tuesday 6 PM · without Liftline
Clipboard ops
- Member queues for squat rack — no idea how long
- Cable machine 3 has a torn handle. For two weeks.
- Staff don't know which cable machine; report said "the broken thing"
- Members leave early. Some never come back.
- Three months later, churn — and the gym never connects it to the cable row.
Tuesday 6 PM · with Liftline
Operational layer
Member sees the squat rack is taken, joins the queue, gets a push when it's their turn — they hit chest in the meantime. The cable machine reports its own issue in seconds; staff get a severity-tagged ticket with photo and machine ID.
Owners get heatmaps of equipment usage, churn-risk scoring correlated to availability, and an AI procurement advisor that turns floor friction into a defensible purchase order.
One platform · three audiences
Each tier sees exactly what they need. Nothing more.
Member app
/app/* · mobile-first
Members open the app, see which equipment is free, tap to join a queue, and get a push when it's their turn. Issue reports take ten seconds — photo plus one sentence, classified by Claude into the right severity tier before staff ever see it.
Floor map
Live equipment status
Queue
Push when it's your turn
NLP
10-second issue reports
Staff dashboard
/staff/* · tablet + desktop
Floor staff see every machine's live status, get triaged maintenance reports ranked by severity, and receive AI anomaly alerts when patterns shift — equipment failure-rate spikes, member-frustration signals, queue buildup that's about to become a complaint.
Triage
Severity-tagged reports
AI
Anomaly alerts / 15min
Realtime
Floor-wide Socket.io
Admin portal
/portal/* · desktop
Owners get utilization heatmaps, churn-risk scoring correlated to equipment availability, an AI procurement advisor that says 'you need two more cable machines — here's the data,' and a full prose-report generator for board-pack-ready monthly summaries.
Heatmaps
Hourly + seasonal
Churn
Retention risk scoring
Reports
AI prose summaries
Three portals · one codebase
Role-based access, shared real-time backbone.
Member app
Mobile-first floor map + queue.
Members see live equipment status, join a queue when a machine is taken, and get notified when it's their turn. Report an issue with one sentence and a photo — Claude triages and classifies on the backend.
Staff dashboard
Tablet-grade floor operations.
Real-time floor view of every machine. Triaged issue queue with severity tags. Cleaning rounds auto-scheduled by usage intensity rather than fixed timers. Anomaly alerts when a member queues unsuccessfully three times this week.
Admin portal
Owner-grade analytics + AI command center.
Heatmaps by hour and day. Churn-risk scoring correlated to equipment availability. Procurement advisor: 'You need two more cable machines — here's the data.' Full AI report generator for utilization, maintenance and financial summaries.
Where AI earns its place
Thirteen Claude-powered features, not a chatbot bolt-on.
Every AI surface answers a specific operational question — wait time, issue severity, churn risk, procurement decision. All cached at the right TTL to keep API costs predictable.
The product in hand
Real screens from the running build.
Not Figma mockups — these are the real portals running against a seeded gym. Charter pilots replace the demo data with yours.
/portal/dashboard
Equipment inventory — status, warranty, maintenance history, cost analysis
/portal/equipment
Live queues — promote, remove, force-end stuck sessions, full audit trail
/portal/queues
Members — tier, status, last visit, derived churn risk
/portal/members
AI Command Center — predictive maintenance, churn, segments, procurement
/portal/ai
Notifications — broadcast by segment, audit trail of every send
/portal/notifications
Floor view — every machine, live status, current member, queue depth
/app/floor (staff role)
Equipment detail — who's on it, who's waiting, how long
/app/floor/[id]
AI equipment insight — failure risk, predicted issue, recommended service
/staff (equipment drilldown)Dark theme shown · light theme available · branding white-labelled per gym
How it's built
Architecture: three portals, one real-time backbone.
Three role-based clients
Member app · Staff dashboard · Admin portal — Next.js 14
Express API + Socket.io
Auth · RBAC · Audit · Rate limiting · Real-time push <100ms
AI service (Anthropic Claude)
13 features · cached 5m–24h · Pydantic-style validated responses
PostgreSQL 16
15 tables · members · equipment · sessions · queues · audit log
Redis cache
Hot queue state · wait-time predictions · AI response cache
Audit log — every decision reproducible
Equipment overrides · member actions · AI dispositions · timestamps
Frontend
Next.js 14 · TypeScript · Tailwind · 3 role-based shells
Backend
Node.js · Express · TypeScript · Auth + RBAC + audit
Database
PostgreSQL 16 · 15 tables · full audit logging
Realtime
Socket.io · <100ms equipment + queue updates
AI
Anthropic Claude · 13 AI features · cached at 5m–24h
Deploy
Containerised · runs in your cloud (AWS / GCP / Azure)
Non-negotiables
- RBAC enforced at the API. Members never see staff data; staff never see admin analytics.
- Real-time updates push to every connected client in under 100 ms via Socket.io.
- Full audit logging on every decision — equipment override, member action, AI disposition.
- AI calls cached at 5 min – 24 h depending on volatility. Predictable API spend.
- Containerised. Runs in your cloud — AWS, GCP, or Azure. Database and Socket.io behind your VPC.
- Member PII stays in your tenant. No SaaS data egress for the AI features.
How it lands
Two-week install. Charter pilot terms.
We install Liftline in your venue free of build cost in exchange for co-design input — your operational data trains the platform's defaults. Limited charter slots for the first small set of operators.
Site survey + equipment ingest
We map your floor plan, equipment inventory and existing tooling (CRM, payments, access control). Tag each machine, photograph the floor, model your peak-hour patterns.
Install + member onboarding
Deploy the three portals in your environment. Onboard members through SMS or your existing app. Train floor staff on triage workflow. Calibrate AI thresholds to your venue.
Tune to your floor
Watch the first two weeks of live data. Tune wait-time predictions, anomaly thresholds and queue logic to how your members actually behave. Adjust notification cadence.
Monthly board pack
AI-generated monthly summary: utilization, maintenance burden, churn-risk cohorts, procurement recommendations. Sized for owner review, not data-team consumption.
You leave with
- · Three portals live in your venue
- · Members onboarded, staff trained
- · AI thresholds tuned to your floor
- · Monthly board-pack AI summary
Runs on your stack
- · Containerised via docker-compose
- · Postgres + Redis in your VPC
- · Anthropic API or AWS Bedrock (Claude)
- · No member-data egress
Who it's for
- · Independent commercial gyms
- · Multi-site fitness chains
- · Boutique studios with equipment queues
- · University + corporate fitness
Next step
Let's pilot Liftline at your gym.
We're selecting a small group of charter operators to install Liftline ahead of public launch — free of build cost in exchange for operational data that trains the platform. If your venue has recurring equipment-availability complaints, growing maintenance backlogs, or churn that you can't fully explain, there's a strong chance we can help.
SEYSO Services Inc. · Toronto, ON · info@seysoservices.com