SkyOdyssey MCP
Enables AI clients to explore cheapest destinations, optimize multi-leg flight itineraries, and reference airport/region data via MCP tools and resources.
README
SkyOdyssey MCP
SkyOdyssey MCP exposes the SkyOdyssey flight engine as a Model Context Protocol server.
It is designed for AI clients (Claude Desktop, MCP Inspector, custom MCP hosts) that need:
- Cheapest destination exploration
- Multi-leg itinerary optimization
- Region/airport reference resources
Features
- MCP tools for flight search and itinerary optimization
- Shared core logic with
SkyOdyssey-CLI(logic.py,airports.py) - Country/airport/airline filters
- Budget-aware pruning (
max_budget) - Direct-flight filtering
- Multi-origin and flexible stay support for 3-leg itineraries
- SQLite caching (
flights_cache.db, TTL 6h) - Retry/backoff behavior for unstable provider responses
- Pydantic validation on itinerary outputs
Tools Exposed
get_flights_on_date
Fetch one-way flights for a specific route/date.
Inputs:
origin(IATA)destination(IATA)date(YYYY-MM-DD)adults(default1)seat_type(defaulteconomy)
get_cheapest_destinations
Explore a region from one origin/date and return cheapest options.
Inputs:
origindateregion(defaultEurope)limit(default15)adultsseat_typeexcluded_countriesexcluded_airportsdirect_onlyinclude_airlines,exclude_airlinesmax_budget
find_cheapest_two_city_itinerary
Find cheapest 3-leg itinerary loops.
Inputs include:
origin(single airport or list)start_datestay_days_1/stay_days_2(fixed int or[min, max])regionlimit_per_legexcluded_countries,excluded_airportsmax_itinerariesforce_different_countriesdirect_onlyreturn_origin(open-jaw)include_airlines,exclude_airlinesmax_budget
Guardrails:
originlist max:5limit_per_legmax:30max_itinerariesmax:100- stay range window max width:
14days (for each ofstay_days_1andstay_days_2)
Resources
airports://regions→ available regionsairports://{region}→ airport codes for a region
Requirements
- Python 3.10+
- Internet access (provider scraping via
fast-flights)
Installation
From SkyOdyssey-MCP:
python -m pip install -r requirements.txt
Run the Server
python server.py
Server transport:
stdio(configured inserver.py)
Claude Desktop Example
Add to claude_desktop_config.json:
{
"mcpServers": {
"skyodyssey": {
"command": "python",
"args": ["C:/absolute/path/to/SkyOdyssey/SkyOdyssey-MCP/server.py"]
}
}
}
Project Structure
server.py— MCP server entrypoint and tool/resource registrationlogic.py— flight search, filtering, optimization, cacheairports.py— regions, airport datasets, exclusion helpersrequirements.txt— runtime dependencies
Troubleshooting
Import error for fast_flights
- Reinstall dependencies:
python -m pip install -r requirements.txt - Ensure your Python environment is the same one used to run
server.py
Empty or partial results
- Provider instability/timeouts can happen; retry run
- Reduce search width (
limit_per_leg) and/or constraints - Relax filters (
direct_only, exclusions, budget)
Cache issues
- Delete cache if needed:
- Windows PowerShell:
Remove-Item flights_cache.db - Windows cmd:
del flights_cache.db
- Windows PowerShell:
License
MIT License.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。