Lattency
A metro map of the best places to get coffee and get online.
Live in London, Nairobi, and San Francisco.
Built for the Nimiq Pay Mini Apps Competition.
The idea
Lattency is a crowdsourced metro map of workable coffee spots — cafés, coworking spaces, hotel lobbies, and anywhere else you can sit down with a coffee and a laptop. Venues are stations. The three lines are speed tiers: Express (≥50 Mbps), Local (10–49 Mbps), and Suspended (<10 Mbps).
The engine works everywhere. The first three boards are London, Nairobi, and San Francisco — switch cities from the nav or deep-link directly to /london, /nairobi, or /sf.
The stack
| Layer | Choice | Why |
|---|
| Backend | Base44 (entities, auth, Deno functions) | Free backend with built-in auth, DB, and hosting — no infra to manage |
| Frontend | Next.js 16.3 on Vercel (App Router, RSC, ISR) | Multi-city SSG, schematic/geographic map, realtime Base44 subscriptions |
| Network | Base44 SDK via @base44/sdk | Serverless-safe — the SDK client is shared across invocations |
| Payments | Nimiq Pay via @nimiq/mini-app-sdk | Contributors earn NIM for verified speed tests; sponsors fund bounties |
What we’re building for the Nimiq Mini Apps Competition
A finished Mini App inside Nimiq Pay where anyone in a live city can:
- Open the map of nearby coffee spots.
- Tap a venue to see live wifi speed + stability.
- Run an in-browser speed test.
- Earn NIM when their verified reading completes a sponsor-funded bounty.
The speed test is the trust mechanism. The NIM reward is the growth mechanism. The map is the product.
Vision: workspace-quality ground truth
We believe the best data about a workspace comes from someone who is physically there. Lattency turns that private experience into a public good:
- Verified speed tests (download, upload, ping, jitter, loss) replace hearsay.
- Objective workspace metadata — power outlets, seating, noise level, table space — replace subjective reviews.
- NIM bounties align incentives between contributors, venues, and sponsors.
We started with cafés. The engine supports cafés, coworking spaces, hotel lobbies, libraries, and hybrid venues. We call them all stations.
The agentic opportunity
AI agents cannot walk into a café and run a speed test. They need a trusted, verified, real-time source of physical-world infrastructure data to make decisions for their human users.
Lattency is a physical-world oracle for the agentic era. We don't build the agent — we build the ground-truth layer that agents consume.
Product principles
- Objective, not subjective — no "coffee quality" ratings. Only verifiable facts.
- Hyper-local first — own one neighbourhood, then expand.
- Crypto-native incentives — bounties are funded and paid in NIM/USDT.
- Contributor-first UX — run a test, earn NIM, repeat.
Project layout
app/
├── page.tsx # Root redirect → /london (carries query params)
├── [city]/page.tsx # Dynamic city route (London / Nairobi / SF), SSG
├── tour/ # Story page + self-running product-loop reel
├── cafes/[slug]/ # Per-venue standalone page
├── partners/ # Sponsorship / bounty dashboard
├── me/ # Contributor profile + payout rail
├── opengraph-image.tsx # OG image (1200×630, "Live in 3 cities")
├── api/cafes/ # POST /api/cafes (create venue + first measurement)
├── api/cafes/near # GET nearby venues (Haversine filter)
├── api/cafes/[id] # GET venue detail + stats
├── api/measurements # POST /api/measurements
├── api/bounties/claim/ # NIM payout flow
└── api/speedtest/ # download.bin, upload sink, whereami
base44/ # Base44 backend config
├── entities/
│ ├── Cafe.json # Venue entity schema
│ ├── Measurement.json # Speed measurement schema
│ ├── Sponsorship.json # Sponsor schema
│ └── Bounty.json # Bounty schema
├── functions/ # 7 Deno backend functions (service role)
└── agents/
└── workspace_concierge.jsonc # AI concierge agent
components/
├── top-nav.tsx # Sticky nav (multi-city links)
├── map-shell.tsx # SVG schematic ↔ Leaflet geographic map, optimistic pins, trail
├── live-map.tsx # Realtime wrapper (Measurement.subscribe)
├── city-switcher.tsx # Split-flap multi-city selector
├── station-directory.tsx # List + filter + geolocation
├── cafe-detail.tsx # Venue detail drawer
├── cafe-page.tsx # Per-venue page
├── cafe-contribution-form.tsx # Add a new venue flow + immediate tier verdict
├── measurement-form.tsx # In-browser speed test
├── bounties-board.tsx # City-aware sponsor-funded bounties
├── concierge-chat.tsx # AI oracle drawer
├── ai-venue-summary.tsx # Auto-generated editorial one-liner
├── map-toast.tsx # Shared toast system (arrival, promotion)
├── onboarding-overlay.tsx # First-visit coach ticket
└── loop-storyboard.tsx # Self-running product reel (SVG animation)
lib/
├── base44.ts # Base44 SDK client singleton
├── base44-data.ts # Data-access layer
├── cafes.ts # Read path + mock fallback
├── cities.ts # City registry (London, Nairobi, SF)
├── cafe-metadata.ts # Workspace metadata vocabulary
├── map-data.ts # Tier paths, waypoints, TIER_USE, tierForDown()
├── measurements.ts # Shared insert path
├── rate-limit.ts # Rate-limiting + outlier logic
├── speedtest.ts # In-browser speed test
├── bounties.ts # City-aware bounty logic
├── mock-cafes.ts # Bundled fallback (33 stations, 3 cities)
└── types.ts # Shared types
hooks/
├── use-personal-trail.ts # localStorage per-city contributed stations
└── use-realtime-cafes.ts # Measurement.subscribe wrapper
Key data model
Venue (Cafe entity)
| Field | Type | Notes |
|---|
name | string | Required |
neighbourhood | string | e.g. "Shoreditch" |
latitude | number | Required |
longitude | number | Required |
venue_type | string | cafe, coworking, hotel-lobby, library, hybrid |
vibe | string | Short editorial descriptor |
city | string | Default "london" |
price_tier | string | budget, mid, premium |
milk_options | array | e.g. ["oat", "soy"] |
power_outlets | boolean | Are outlets available? |
seating | string | bar, tables, lounge, mixed |
noise_level | string | quiet, moderate, loud |
table_space | string | small, standard, large |
wifi_network | string | SSID or password hint |
photo_url | string | Base64 data URL |
Measurement
| Field | Type | Notes |
|---|
cafe_id | string | Required — Venue entity ID |
down_mbps | number | Required |
up_mbps | number | Required |
latency_ms | number | Required |
jitter_ms | number | |
loss_pct | number | |
measured_at | datetime | |
time_bucket | string | morning, afternoon, evening |
photo_url | string | |
test_method | string | manual, browser-auto |
target_server | string | Edge region id |
device_type | string | Derived from User-Agent |
contributor_ip_hash | string | Hashed for rate-limiting only |
is_outlier | boolean | Flagged by stats function |
contributor_user_id | string | Base44 User ID |
Nimiq integration
Contributors earn NIM for verified speed tests that complete sponsor-funded bounties. The flow:
- Sponsor funds a bounty in NIM (e.g., "3 verified tests at oat-milk cafés in Shoreditch").
- Contributor opens the Mini App, finds a venue matching the bounty.
- Contributor runs the in-browser speed test.
- On a verified, non-outlier reading, the bounty pays NIM directly to the contributor’s Nimiq Pay wallet.
NIM is the default reward token. USDT on supported EVM chains can be added later.
Quick start
pnpm install
npx base44 create # scaffold or link the Base44 backend
npx base44 entities push # deploy entity schemas
npx base44 functions deploy # deploy geo-filter + stats functions
pnpm dev # random high port (see AGENTS.md)
Development
Type-check the whole project:
pnpm exec tsc --noEmit
Run the smoke suite:
pnpm test
Deploy
Frontend
pnpm build
vercel --prod
Backend
npx base44 deploy
License
MIT
