Aerobase Travel Concierge

⚠️ TRANSPARENCY NOTICE

This skill creates a small JSON file in your home directory for rate limiting:

• ~/aerobase-browser-searches.json - Browser search count (max 10/day)

These files contain only:

• Date and count of operations
• Site queried (for browser searches)

Data handling:

• Gmail scanning only happens when user explicitly requests it
• No personal data is sent to external services beyond what's required

External Services:

Scrapling is a stealth browsing service for price verification and deal discovery. It bypasses anti-bot systems to fetch live prices from Google Flights, Kayak, etc.

Aerobase Travel Concierge ⭐ ALL-IN-ONE

Your personal AI travel agent that never sleeps. This is the complete package — all Aerobase skills in one installation.

API-First Principle

ALWAYS use the Aerobase API first before any browser automation. The API is faster, more reliable, and doesn't trigger anti-bot systems.

Browser/scraping is only for:

• Verification of live prices
• Sites not covered by API
• Check-in automation (experimental)
• User explicitly requests external site check

✈️ Flight Search & Booking

API Endpoints

GET /api/v1/flights/search — Search flights with jetlag scoring POST /api/v1/flights/score — Score any flight option (0-100 jetlag) POST /api/v1/flights/compare — Compare multiple options

Query params: origin, destination, departure, return, cabin, maxPrice, limit

Example

POST /api/v1/flights/search
{ "origin": "JFK", "destination": "NRT", "departure": "2026-03-15", "cabin": "business" }

Returns flights with jetlag scores, recovery days, and recommendations.

🏆 Award Search

API Endpoints

POST /api/v1/awards/search — Search 24+ airline loyalty programs GET /api/transfer-bonuses — Current transfer partner bonuses

Query: origin, destination, departure, return, cabin, maxMiles

Programs Supported

Aeromexico, Air Canada, Alaska, Alitalia, All Nippon, American, Avianca, British Airways, Cathay Pacific, Delta, Emirates, Etihad, Flying Blue, Iberia, Japan Airlines, JetBlue, Korean Air, Lufthansa, Qantas, Qatar, Singapore, Southwest, United, Virgin Atlantic

CPP Calculation

Use Scrapling to get cash prices for cents-per-point calculation:

POST {SCRAPLING_URL}/search
{"site":"google-flights","origin":"JFK","destination":"LHR","departure":"2026-04-01","return":"2026-04-08"}

Reference: Scrapling Docs

🏧 Airport Lounges

API Endpoints

GET /api/v1/lounges — Search with filters: airport, airline, network, tier, search GET /api/airports/{code}/lounges — Lounges at specific airport

Data sourced from LR (LoungeReview) tables with detailed information.

Jetlag-Relevant Fields

• recoveryScore: 1-10, higher = better for recovery
• hasShowers: Boolean - freshen up between flights
• hasSpa: Boolean - premium recovery
• hasSleepPods: Boolean - rest between flights
• quietZone: Boolean - circadian alignment
• naturalLight: Boolean - jetlag adjustment

Scrapling — Priority Pass Hours

Use for real-time hours verification:

web_fetch {SCRAPLING_URL}/fetch?url=https://www.prioritypass.com/lounges/united-states/new-york-ny/jfk-john-f-kennedy-intl&json=1&extract=css&selector=.lounge-card

Reference: Scrapling Docs

🏨 Hotel Search

API Endpoints

GET /api/v1/hotels — Search with filters: airport, city, country, chain, tier, stars, jetlagFriendly, search GET /api/hotels/near-airport/{code} — Airport-adjacent hotels

Day-Use Hotels (Layovers)

GET /api/dayuse?airport={code} — Day-use hotels for layover passengers (book by the day, no overnight) GET /api/hotels?dayuse=true — Filter hotels to day-use only

For layovers > 8 hours, recommend day-use hotels.

🎫 Tours & Activities

API Endpoints

GET /api/v1/tours — Viator tours and attractions Query: airport, destination, category, duration, priceRange

Scrapling — TripAdvisor Discovery

Use for supplementary activity discovery:

web_fetch {SCRAPLING_URL}/fetch?url=https://www.tripadvisor.com/Attractions-g294217-Activities-Tokyo_Tokyo_Prefecture_Kanto.html&json=1&extract=css&selector=.listing_title

Reference: Scrapling Docs

💰 Flight Deals

API Endpoint

GET /api/v1/deals — Curated flight deals Query: origin, destination, maxPrice, minScore, dateFrom, dateTo, cabin, sort, limit

Returns up to 50 results (3 for free tier).

Data Sources

The Deals API includes deals scraped from:

• SecretFlying
• TheFlightDeal
• TravelPirates
• Going.com

These sources are scraped periodically and stored in our database.

Scrapling — Deal Verification

Only use for verification or when user explicitly asks:

POST {SCRAPLING_URL}/search
{"site":"secretflying"}
POST {SCRAPLING_URL}/search
{"site":"google-flights","origin":"JFK","destination":"NRT","departure":"2026-03-15","return":"2026-03-22"}

Reference: Scrapling Docs

💳 Points & Wallet

API Endpoints

GET /api/v1/credit-cards — Credit card transfer partners GET /api/v1/loyalty/balances — (Premium) Gmail scanning required

Gmail Loyalty Scanning (Premium)

⚠️ User Consent Required: Only use this feature when:

1. User explicitly asks "scan my Gmail for points" or "show my loyalty balances"
2. User has already connected Gmail via OAuth in Aerobase settings (handled by Aerobase, not this skill)
3. Never scan Gmail proactively - only after user request

How it works:

• Users connect Gmail via OAuth in Aerobase settings. Only loyalty/points data is extracted - email contents are never shared.
• The skill calls GET /api/v1/loyalty/balances - Aerobase server handles the Gmail API call
• Only points/loyalty numbers are returned - never email content

😴 Jetlag Recovery Plans (PREMIUM)

API Endpoint

POST /api/v1/recovery/plan — Generate personalized recovery plan

Requires: flight (origin, destination, departureTime), chronotype (earlyBird, nightOwl, neutral), tripDuration

Returns:

• Pre-trip preparation schedule
• In-flight strategies
• Day-by-day recovery timeline
• Light exposure timing
• Sleep and diet adjustments

🔧 Browser Automation

USE BROWSER ONLY WHEN:

• User specifically asks to check Google Flights / Kayak / Skyscanner
• API search returned no results and user wants broader coverage
• Price comparison requested
• Check-in automation (experimental)

Browser Commands (OpenClaw Playwright-on-CDP)

• browser snapshot — get ARIA tree with [ref=eN] element references
• browser type [ref=eN] "value" — type into input field
• browser click [ref=eN] — click element
• browser screenshot — capture current page state

Context Selection

Scrapling Service (Anti-Bot Bypass)

When browser automation is blocked by anti-bot systems (Akamai, Cloudflare, Datadome), use the stealth scrapling service configured via SCRAPLING_URL environment variable.

Reference: Scrapling Documentation

Fetch a page

web_fetch {SCRAPLING_URL}/fetch?url=https://www.delta.com&json=1

Run JavaScript

POST {SCRAPLING_URL}/evaluate
Body: {"url": "https://seats.aero", "script": "document.title"}

Multi-step interaction (forms)

POST {SCRAPLING_URL}/interact
{
  "url": "https://www.example.com/form/",
  "steps": [
    {"action": "consent"},
    {"action": "fill", "selector": "#confirmationNumber", "value": "<PNR>"},
    {"action": "fill", "selector": "#passengerLastName", "value": "<LAST>"},
    {"action": "click", "selector": "button#form-mixin--submit-button"},
    {"action": "wait", "ms": 5000},
    {"action": "screenshot"}
  ]
}

Aggregator Search

POST {SCRAPLING_URL}/search
{"site":"google-flights","origin":"LAX","destination":"NRT","departure":"2026-03-15","return":"2026-03-22"}

Rate Limits

• Browser searches: max 10/day per user
• Scrapling: follow cache rules (5-min cache)
• API: varies by endpoint (see each section)

📋 Rate Limits Summary

Combined API calls: 5/day total (all endpoints share the same quota)

📁 Additional Files

Helper Scripts (scripts/)

• browser-tracking.mjs - Track browser search limits

API References (references/)

• aerobase-api.md - Complete Aerobase API endpoint reference
• scrapling.md - Scrapling service endpoints and parameters

Full Documentation

• OpenAPI Spec: https://aerobase.app/api/v1/openapi
• Scrapling Docs: https://scrapling.readthedocs.io/en/latest/overview.html

🔒 Privacy, Security & Credentials

Required Environment Variables

Note: SCRAPLING_URL is configured by the OpenClaw deployment - users do not need to provide this.

Gmail OAuth - How It Works

1. User navigates to https://aerobase.app/settings
2. Clicks "Connect Gmail" - goes to Google's OAuth flow
3. User grants read-only access to loyalty/points emails only
4. Aerobase receives OAuth token - token is stored by Aerobase, NOT by this skill
5. Skill calls GET /api/v1/loyalty/balances - Aerobase server makes the Gmail API call
6. Email contents never leave Aerobase servers - only extracted points balances returned

What data is extracted:

• Loyalty program account numbers (from airline/hotel point statements)
• Point balances mentioned in emails
• Credit card spend categories

What is NOT accessed:

• Email body/content
• Personal messages
• Contacts
• Calendar

Local File Storage

The skill creates a small JSON file in the user's home directory to track rate limits:

• ~/aerobase-browser-searches.json - Browser search count and history

Contents stored:

• Date and count of browser searches
• Site queried and timestamp

These files are used only for rate limiting - no personal data beyond search counts.

User control:

• User can revoke access anytime from Google's account settings
• Aerobase deletes tokens on disconnect

See: https://aerobase.app/privacy

Scrapling Service

The {SCRAPLING_URL} service is operated by the OpenClaw deployment. It is used for:

• Price verification (Google Flights, Kayak)
• Deal site checks (SecretFlying, TheFlightDeal)

Data sent to Scrapling:

• URLs requested by the skill (price verification, deal discovery)
• No user credentials or PII

Not used for:

• Gmail access
• Payment information
• Storing user data

Browser Automation

For live price verification:

• Fetches current prices from travel sites
• No user credentials required

Security & Privacy Links:

• Aerobase Privacy Policy: https://aerobase.app/privacy
• Aerobase Terms: https://aerobase.app/terms
• OpenClaw Security: https://openclaw.ai/security

🚀 Quick Reference

Always

1. Query API first - faster, reliable, no blocks
2. Show API results to user immediately
3. Use browser/scrapling only for verification or explicit requests
4. Enrich results with jetlag scores

Fallback Chain

1. Try API first
2. Try Scrapling service (no proxy needed)
3. Try native browser + proxy
4. Screenshot and explain to user if all fail