Serpapi

---
name: serpapi
description: Google Flights cash prices, Google Hotels, and Google Travel Explore via SerpAPI. Use for award-vs-cash comparison, hotel search, and destination discovery.
category: hotels
summary: Google Hotels search and destination discovery.
api_key: SerpAPI
license: MIT
---

# SerpAPI Skill

Scrape Google Flights, Google Hotels, and Google Travel Explore via SerpAPI. Provides cash flight prices (for Chase/Amex portal comparison), hotel pricing, and destination discovery.

**Source:** [serpapi.com](https://serpapi.com) — Free tier available, paid plans for higher volume.

## Authentication

`SERPAPI_API_KEY` is set in `.env`. All requests use `api_key` query parameter.

## API Base

```
https://serpapi.com/search
```

## Google Flights (Cash Prices)

Search for flight prices and schedules. Essential for comparing: "Is 88,000 United miles better than paying $900 cash through the Chase portal?" (Chase portal pricing is now dynamic via Points Boost, ~1.5-2.0 cpp on select bookings; verify the actual quote.)

### One-Way Search

```bash
curl -s "https://serpapi.com/search?engine=google_flights&departure_id=SFO&arrival_id=NRT&outbound_date=2026-08-10&type=2&adults=2&travel_class=1&currency=USD&stops=2&sort_by=2&api_key=$SERPAPI_API_KEY" | jq '{best: [.best_flights[]? | {price: .price, duration: .total_duration, stops: (.layovers | length), flights: [.flights[] | {from: .departure_airport.id, to: .arrival_airport.id, airline: .airline, flight: .flight_number, depart: .departure_airport.time, arrive: .arrival_airport.time}]}], price_insights: .price_insights}'
```

### Parameters

| Param | Required | Description |
|-------|----------|-------------|
| `engine` | Yes | `google_flights` |
| `departure_id` | Yes | Airport code(s), comma-separated: `SFO,PDX` |
| `arrival_id` | Yes | Airport code(s), comma-separated: `NRT,HND` |
| `outbound_date` | Yes | `YYYY-MM-DD` |
| `return_date` | Round trip | `YYYY-MM-DD` (required if type=1) |
| `type` | No | `1` = round trip (default), `2` = one way, `3` = multi-city |
| `adults` | No | Default 1 |
| `children` | No | Default 0 |
| `travel_class` | No | `1` = economy, `2` = premium economy, `3` = business, `4` = first |
| `stops` | No | `0` = any, `1` = nonstop, `2` = 1 stop or fewer, `3` = 2 stops or fewer |
| `sort_by` | No | `1` = top flights, `2` = price, `3` = departure, `4` = arrival, `5` = duration |
| `include_airlines` | No | IATA codes: `SK,KL,UA` or alliances: `STAR_ALLIANCE,SKYTEAM,ONEWORLD` |
| `max_price` | No | Maximum ticket price in USD |
| `max_duration` | No | Maximum flight duration in minutes |
| `bags` | No | Number of carry-on bags |
| `deep_search` | No | `true` for browser-identical results (slower) |
| `currency` | No | Default `USD` |

### Multi-City (Open Jaw)

Use `type=3` with `multi_city_json`:

```bash
curl -s "https://serpapi.com/search?engine=google_flights&type=3&multi_city_json=%5B%7B%22departure_id%22%3A%22SFO%22%2C%22arrival_id%22%3A%22NRT%22%2C%22date%22%3A%222026-08-05%22%7D%2C%7B%22departure_id%22%3A%22ICN%22%2C%22arrival_id%22%3A%22SFO%22%2C%22date%22%3A%222026-08-26%22%7D%5D&adults=2&travel_class=1&currency=USD&api_key=$SERPAPI_API_KEY" | jq '.'
```

The JSON value for multi_city_json is URL-encoded. Decoded:
```json
[{"departure_id":"SFO","arrival_id":"NRT","date":"2026-08-05"},{"departure_id":"ICN","arrival_id":"SFO","date":"2026-08-26"}]
```

### Response Fields

Each flight in `best_flights[]` and `other_flights[]`:

| Field | Description |
|-------|-------------|
| `price` | Cash price in USD |
| `total_duration` | Total minutes |
| `flights[]` | Array of legs with airline, flight number, times, airplane, legroom |
| `layovers[]` | Array with duration and airport for each connection |
| `departure_token` | Token to get return flight options (round trip) |
| `booking_token` | Token to get booking options |

`price_insights` includes `lowest_price`, `price_level` (low/typical/high), and `typical_price_range`.

### Portal Comparison Math

Chase Sapphire Reserve: dynamic Points Boost pricing, typically 1.5-2.0 cpp on select bookings (not a fixed floor). Verify actual portal price for the specific booking.
If cash price is $900, portal cost = 60,000 UR points.
If award price is 88,000 United miles, cash via portal is better value.

Amex: typically 1 cpp via portal (worse value, use transfers instead).

Capital One Venture X: 1 cpp via portal, but transfer partners can be better.

## Google Hotels

Search hotels and vacation rentals with pricing from multiple OTAs.

```bash
curl -s "https://serpapi.com/search?engine=google_hotels&q=hotels+Tokyo+Japan&check_in_date=2026-08-10&check_out_date=2026-08-13&adults=2&currency=USD&sort_by=3&api_key=$SERPAPI_API_KEY" | jq '[.properties[]? | {name: .name, type: .type, rating: .overall_rating, reviews: .reviews, price: .rate_per_night.extracted_lowest, class: .extracted_hotel_class, amenities: .amenities}] | .[0:10]'
```

### Parameters

| Param | Required | Description |
|-------|----------|-------------|
| `engine` | Yes | `google_hotels` |
| `q` | Yes | Search query: `hotels Tokyo Japan` |
| `check_in_date` | Yes | `YYYY-MM-DD` |
| `check_out_date` | Yes | `YYYY-MM-DD` |
| `adults` | No | Default 2 |
| `children` | No | Default 0 |
| `sort_by` | No | `3` = lowest price, `8` = highest rating, `13` = most reviewed |
| `min_price` / `max_price` | No | Price range filter |
| `hotel_class` | No | `2,3,4,5` (comma-separated) |
| `rating` | No | `7` = 3.5+, `8` = 4.0+, `9` = 4.5+ |
| `vacation_rentals` | No | Set to `true` for Airbnb-style results |
| `property_token` | No | Get details for a specific property |

## Google Travel Explore

Discover destinations and cheapest flights from an origin. Great for "where can I fly cheaply in August?"

```bash
curl -s "https://serpapi.com/search?engine=google_travel_explore&departure_id=SFO&outbound_date=2026-08-05&return_date=2026-08-26&adults=2&travel_class=1&currency=USD&api_key=$SERPAPI_API_KEY" | jq '[.destinations[]? | {name: .name, country: .country, airport: .destination_airport.code, price: .flight_price, duration: .flight_duration, stops: .number_of_stops, airline: .airline}] | .[0:15]'
```

### Parameters

| Param | Required | Description |
|-------|----------|-------------|
| `engine` | Yes | `google_travel_explore` |
| `departure_id` | Yes | Airport code or kgmid |
| `arrival_id` | No | Specific destination |
| `arrival_area_id` | No | Region kgmid (e.g., `/m/02j9z` for Europe) |
| `outbound_date` | No | `YYYY-MM-DD` |
| `return_date` | No | `YYYY-MM-DD` |
| `month` | No | `1`-`12` for flexible dates |
| `travel_duration` | No | `1` = weekend, `2` = 1 week, `3` = 2 weeks |
| `interest` | No | `/g/11bc58l13w` = Outdoors, `/m/0b3yr` = Beaches |
| `include_airlines` | No | Filter by airline or alliance |
| `max_price` | No | Maximum price |
| `stops` | No | Same as Google Flights |

## Workflow: Compare Award vs Cash

1. Search cash prices on Google Flights via SerpAPI
2. Estimate portal cost. Chase uses dynamic "Points Boost" pricing (~1.5-2.0cpp on select bookings, not a flat rate). Amex/Capital One ~1.0cpp. For rough math, run the actual portal quote against the cash price; do not assume a flat cpp on Chase.
3. Compare with award price from Seats.aero
4. Lower number wins (accounting for the value you place on each currency)

## Notes

- Cached results are free (1hr cache). Set `no_cache=true` to force fresh.
- `deep_search=true` gives browser-identical results but is slower.
- Results include `price_insights` with historical price data and trend.
- Multi-city supports open jaw itineraries natively.
- Hotels support vacation rentals mode for Airbnb-style results.