什么是 Google 航班搜索？

Google 航班搜索技能允许您的 AI 智能体直接从 Google Flights 获取实时航班时刻表和定价。通过利用 SearchAPI.io 接口，它为开发者提供了一种稳健的方式，将全球航班搜索集成到其 Openclaw Skills 生态系统中。该工具不仅限于简单的数据检索，还集成了评分和坚控工作流，以确保用户获得最佳的旅行选择。

该技能旨在轻量高效，利用标准 Python 库与外部 API 交互。它被设计为多技能流水线的一部分，确保每次搜索不仅向用户显示，还会保存用于持续价格跟踪，使其成为任何面向旅游的 Openclaw Skills 实现中不可或缺的组件。

下载入口:https://github.com/openclaw/skills/tree/main/skills/galdayan/google-flights-search

安装与下载

1. ClawHub CLI

从源直接安装技能的最快方式。

2. 手动安装

将技能文件夹复制到以下位置之一

优先级：工作区 > 本地 > 内置

3. 提示词安装

将此提示词复制到 OpenClaw 即可自动安装。

Install

Google 航班搜索 应用场景

• 使用 IATA 代码查找特定或灵活日期范围内的最便宜航班。
• 通过自动跟踪坚控特定航线随时间的价格下降。
• 比较往返和单程选项，包含详细的经停和持续时间信息。
• 根据经停次数、舱位等级或特定航空公司偏好过滤航班。
• 自动发现航空公司和在线旅行社的直接预订链接。

1. 从用户查询中提取出发地、目的地和日期等航班参数。
2. 执行 search_searchapi.py 脚本，通过 SearchAPI.io 从 Google Flights 提取实时数据。
3. 将结果传递给航班评分技能，按价值和便利性对选项进行排名。
4. 使用唯一预订令牌获取评分最高结果的直接预订链接。
5. 使用航班结果格式化工具将最终数据格式化为易于阅读的表格。
6. 自动将搜索注册到航班价格坚控器，以便进行长期跟踪和通知。

Google 航班搜索 配置指南

首先，使用 hub 命令安装该技能：

clawhub install google-flights-search

接下来，确保您拥有 SearchAPI.io 账号和密钥。将您的密钥添加到项目根目录下的 .env 文件中：

SEARCHAPI_KEY=your_api_key_here

Openclaw Skills 将自动加载此环境变量。由于该技能使用 Python 3 标准库，因此不需要额外的 pip 安装。

Google 航班搜索 数据架构与分类体系

该技能返回一个包含航班详情和元数据的结构化 JSON 对象。以下是主要结构：

name: google-flights-search
description: "Search real flight prices and schedules from Google Flights via SearchAPI.io. Use when a user asks to find flights, check prices, compare options, or search around a date range. Always pipe results through the flight-scoring skill to rank before presenting, then ALWAYS save the search via the flight-price-monitor skill for automatic price tracking. Requires SEARCHAPI_KEY in .env."
metadata: {"requires": {"env": ["SEARCHAPI_KEY"]}}

Google Flights Search

Fetch live flight data from Google Flights via SearchAPI.io, then score and rank results using the flight-scoring skill, then ALWAYS save the search via the flight-price-monitor skill for automatic price tracking.

Install

clawhub install google-flights-search

Requirements

• Python 3 — uses only stdlib (urllib, json, argparse). No pip installs needed.
• SearchAPI.io account — free tier includes 100 requests/month. Sign up here.
• SEARCHAPI_KEY — get your API key from SearchAPI.io after registering, then add it to .env at the project root. OpenClaw loads it automatically.

Quick Reference

Usage

# One-way search
python {baseDir}/scripts/search_searchapi.py r
  --from TLV --to LON --date 2026-03-15 --days 3 --currency USD

# Round-trip with return flight details for top 5 (RECOMMENDED for round-trips)
python {baseDir}/scripts/search_searchapi.py r
  --from TLV --to BKK --date 2026-03-28 --return-date 2026-04-14 --top 5

All Parameters

Destination codes

Google Flights accepts:

• Airport codes: LHR, CDG, TLV
• City codes: LON (all London airports), PAR (all Paris), NYC

Use city codes when the user hasn't specified a preferred airport.

Output Format

The script prints JSON to stdout.

One-Way / Outbound Only

{
  "origin": "TLV",
  "destination": "LON",
  "flight_type": "one_way",
  "date_range": { "from": "2026-03-15", "to": "2026-03-17", "days_searched": 3 },
  "currency": "USD",
  "total_results": 24,
  "showing": 15,
  "flights": [
    {
      "search_date": "2026-03-15",
      "airline": "El Al",
      "flight_number": "LY315",
      "origin": "TLV",
      "destination": "LHR",
      "departure_time": "22:00",
      "departure_date": "2026-03-15",
      "arrival_time": "01:30",
      "arrival_date": "2026-03-16",
      "duration_minutes": 270,
      "stops": 0,
      "layovers": [],
      "min_layover_minutes": null,
      "price": 1850,
      "overnight": true,
      "booking_token": "WyJDa..."
    }
  ]
}

Round-Trip with --top 5 (includes return flight details)

When --top N is used with --return-date, each of the top N outbound results includes a nested return_flight object with full return leg details:

{
  "origin": "TLV",
  "destination": "BKK",
  "flight_type": "round_trip",
  "return_date": "2026-04-14",
  "return_flights_fetched_for_top": 5,
  "flights": [
    {
      "search_date": "2026-03-28",
      "airline": "Etihad",
      "flight_number": "EY 610",
      "origin": "TLV",
      "destination": "BKK",
      "departure_time": "07:20",
      "departure_date": "2026-03-28",
      "arrival_time": "23:25",
      "arrival_date": "2026-03-28",
      "duration_minutes": 725,
      "stops": 1,
      "layovers": [{"airport": "Zayed International Airport", "duration_minutes": 185}],
      "min_layover_minutes": 185,
      "price": 1569,
      "overnight": false,
      "return_flight": {
        "airline": "Etihad",
        "flight_number": "EY 401",
        "origin": "BKK",
        "destination": "TLV",
        "departure_time": "01:05",
        "departure_date": "2026-04-14",
        "arrival_time": "08:15",
        "arrival_date": "2026-04-14",
        "duration_minutes": 670,
        "stops": 1,
        "layovers": [{"airport": "Zayed International Airport", "duration_minutes": 150}],
        "min_layover_minutes": 150,
        "overnight": false,
        "price": 1569,
        "booking_token": "EhkIAh..."
      }
    }
  ]
}

Important: The price is the combined round-trip price (same on both outbound and return_flight). The return_flight object has its own airline, flight number, times, layovers, and stops — use all of these when formatting results.

Booking Token Lookup (--booking-token)

After scoring, use the booking_token from each top result to fetch real booking URLs.

You must pass the same --from, --to, --date used in the original search. For round-trips, also pass --return-date.

# One-way booking lookup
python {baseDir}/scripts/search_searchapi.py r
  --from TLV --to LON --date 2026-03-15 r
  --booking-token "WyJDa..."

# Round-trip booking lookup — MUST include --return-date
python {baseDir}/scripts/search_searchapi.py r
  --from TLV --to SGN --date 2026-03-29 --return-date 2026-04-14 r
  --booking-token "WyJDa..."

Returns:

{
  "mode": "booking_lookup",
  "total_options": 3,
  "booking_options": [
    {
      "book_with": "El Al",
      "price": 1850,
      "url": "https://www.elal.com/..."
    },
    {
      "book_with": "Kiwi.com",
      "price": 1820,
      "url": "https://www.kiwi.com/..."
    }
  ]
}

Use the first option whose book_with matches the airline (direct airline link). If no airline match, use the first option. If --booking-token fails or returns no options, fall back to vendor-booking-link.

Full Workflow

1. User requests flights
        ↓
2. Extract: origin, destination, date(s), passengers
        ↓
3. Run search_searchapi.py with --top 5 for round-trips
   → each result includes outbound + nested return_flight details + booking_token
        ↓
4. Apply flight-scoring rules to each result:
   - Calculate price score (0–50)
   - Calculate direct score (0–30)
   - Calculate convenience score (0–20)
   - Sum → total score
        ↓
5. Assign tags (Best overall, Cheapest, Direct, Night flight, +1 not a working day)
        ↓
6. Get booking links for each top result:
   a. Use booking_token → run search_searchapi.py --booking-token TOKEN → real airline/OTA URLs
   b. Fallback: if booking_token missing or fetch fails → use vendor-booking-link skill
        ↓
7. Format results using flight-results-formatter → table with both legs per option
        ↓
8. Present the formatted table to user
        ↓
9. Run save_monitor.py → save search for price tracking (MANDATORY)

Always score before presenting. Never dump raw JSON to the user. Always include booking links. Prefer booking_token for real URLs; fall back to vendor-booking-link if unavailable. Never link to Google Flights or aggregators. Round-trips MUST use --top 5 to get return flight details. Without --top, only outbound info is returned.

Step 8: Save Price Monitor (MANDATORY)

After presenting results, ALWAYS save the search for price monitoring. Build a compact snapshot from the top 10-15 flights as "Airline|FlightNum|DepTime": price pairs, then run:

python3 {baseDir}/../flight-price-monitor/scripts/save_monitor.py r
  --user-id "" r
  --from  --to  --date  r
  --return-date  r
  --currency  --adults  r
  --channel  --delivery-to "" r
  --flights ''

• --user-id: The peer ID from the session (e.g. whatsapp:+972523866782). If unknown, use the user's name.
• --channel: The channel the user is on (whatsapp, telegram, etc.)
• --delivery-to: The user's address on that channel. Use last if unknown.
• --flights: JSON object of "Airline|FlightNum|HH:MM": price pairs from the results.
• Omit --return-date if it was a one-way search.

This overwrites any previous monitor for this user. A cron job checks every 2 hours and notifies the user of price changes.

Mapping Script Output → Scoring Inputs

Error Handling

Notes

• El Al does not fly on Shabbat — Friday evening and Saturday flights from TLV won't appear for El Al.
• Secondary airports: Wizz/Ryanair use LTN, STN, BGY etc. — when the script returns results from these, mention the airport to the user.
• Prices are per-person unless --adults is set. For groups, multiply accordingly and note total.
• Round-trip: Use --return-date YYYY-MM-DD --top 5 to search round-trip and auto-fetch return flight details for the top 5 results.