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:

1. Open the map of nearby coffee spots.
2. Tap a venue to see live wifi speed + stability.
3. Run an in-browser speed test.
4. 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

1. Objective, not subjective — no "coffee quality" ratings. Only verifiable facts.
2. Hyper-local first — own one neighbourhood, then expand.
3. Crypto-native incentives — bounties are funded and paid in NIM/USDT.
4. 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:

1. Sponsor funds a bounty in NIM (e.g., "3 verified tests at oat-milk cafés in Shoreditch").
2. Contributor opens the Mini App, finds a venue matching the bounty.
3. Contributor runs the in-browser speed test.
4. 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