Skip to main content
Glama
Sign In
deenesjakoruzh
nxGnosis

Travel Agent MCP Server

by nxGnosis
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Remote

TravelAgentMCP 🌍✈️

Ein Monorepo von MCP-Servern (Model Context Protocol) für die TVA-OTA-Plattform: Flug- und Hotelsuche & -buchung, Visum- und Einwanderungsanforderungen & Buchungsverfolgung sowie Benutzerkontoverwaltung — entwickelt für KI-Reiseagenten.

Architektur

TravelAgentMCP/
├── mcps/
│   ├── shared/            @travelagent-mcp/shared — HTTP client, config, guardrails, tool helper (not a server)
│   ├── flight-mcp/        tva-flight-mcp   — search, pricing, seatmaps, booking lifecycle
│   ├── hotel-mcp/         tva-hotel-mcp    — search, availability, booking lifecycle
│   ├── visa-mcp/          tva-visa-mcp     — country requirements + booking tracking
│   ├── immigration-mcp/   tva-immigration-mcp — service info + booking tracking
│   └── account-mcp/       tva-account-mcp  — auth, profile, promo codes, notifications
└── src/index.ts           travelagent-mcp  — gateway that mounts all five servers' tools in one process

Jedes mcps/*-Paket ist ein unabhängig ausführbarer stdio MCP-Server mit eigenem bin. Führen Sie die benötigten aus (z. B. nur tva-flight-mcp + tva-hotel-mcp für ein Buchungs-Widget oder nur tva-visa-mcp + tva-immigration-mcp für einen Beratungs-Bot) oder führen Sie das root Gateway (travelagent-mcp) aus, um alles in einem Prozess zu erhalten — das ist es, was das veröffentlichte npm-Paket immer war, aus Gründen der Abwärtskompatibilität beibehalten.

Eine vorgelagerte API, zwei Authentifizierungsmodi:

  • TVA OTA Backend (TVA_BASE_URL) — Flüge, Hotels, Buchungen, Konto und die „Client“-Seite der Visum-/Einwanderungs- Buchungsverfolgung (Updates, Benachrichtigungen, Transaktionen). Die meisten Buchungsverwaltungs-Endpunkte erfordern ein Bearer-Token; Gast-Checkout-Abläufe (Buchen, Stornieren nach UUID, Verifizieren, Ticket erneut senden) jedoch nicht.

  • Visum-/Einwanderungsinhalt (gleiche TVA_BASE_URL, authentifiziert über CONTENT_API_KEY anstelle eines Bearer-Tokens) — Länderanforderungen, Visumarten, Gebühren und FAQs, die nur von GET_VISA_INFO_BY_COUNTRY / GET_IMMIGRATION_INFO_BY_COUNTRY verwendet werden.

Related MCP server: orizn-visa-mcp

Schutzmaßnahmen

  • Prinzip der geringsten Rechte: Es werden nur Endpunkte für Benutzer freigegeben. Admin-/Betriebs-Endpunkte (Buchungsverwaltung, Rückerstattungen, Umsatzstatistiken) werden bewusst nicht als Tools bereitgestellt — sie sind nicht für einen Konversationsagenten geeignet.

  • Explizite Bestätigung: Jedes Tool, das Geld bewegt oder irreversibel ist (BOOK_FLIGHT, BOOK_HOTEL, CANCEL_FLIGHT_BOOKING, CANCEL_HOTEL_BOOKING, CONFIRM_FLIGHT_PAYMENT, CONFIRM_HOTEL_PAYMENT, SELECT_FLIGHT_SEAT, CHANGE_PASSWORD, CLOSE_ACCOUNT), erfordert confirm: true, das der aufrufende Agent nur setzen sollte, nachdem er dem Benutzer die spezifischen Details vorgelesen hat.

  • Keine Geheimnisse im LLM-Kontext: TVA_ACCESS_TOKEN ist nie ein Tool-Parameter — es wird aus der Prozessumgebung ausgelesen, pro Benutzersitzung von dem Agenten injiziert, der diese Server startet. LOGIN_USER / SOCIAL_AUTH_LOGIN geben das Token zurück, eingewickelt in einen __session-Umschlag, den der startende Agent abfangen und out-of-band speichern kann; das Modell sieht nur eine einfache Bestätigungsnachricht. Siehe mcps/account-mcp/src/lib/session.ts für den Vertrag.

  • Keine Wiedergabe von Rohangeboten: Amadeus-Flugangebote sind große undurchsichtige JSON-Blobs. SEARCH_FLIGHTS speichert jedes Angebot server-seitig zwischen (SQLite, siehe mcps/flight-mcp/src/lib/offerCache.ts) und übergibt dem Modell stattdessen eine kurze offerRef — das Modell rekonstruiert niemals von Hand Angebots-JSON, was das Risiko von Token-Aufblähung oder stiller Preis-/Tarif-Änderung birgt.

  • Schwärzung: Passwörter, Token, OTPs und Kartennummern werden maskiert, bevor etwas protokolliert wird (siehe mcps/shared/src/guardrails.ts).

  • Bekanntes Risiko — Hotelzahlungen: BOOK_HOTEL akzeptiert derzeit rohe Kartendaten direkt (die TVA-Hotel-API hat noch keinen tokenisierten Checkout-Schritt). Dies darf nur von einem Client aufgerufen werden, der Kartendaten über eine sichere, PCI-konforme Oberfläche gesammelt hat — niemals, indem der Benutzer aufgefordert wird, eine Kartennummer in den Chat einzugeben. Siehe die Beschreibung des Tools für Details.

  • Ratenbegrenzung: Jeder Tool-Aufruf wird prozessintern gedrosselt (Abwehr gegen eine außer Kontrolle geratene Funktionsaufruf-Schleife). Dies ist kein Ersatz für eine persistente, sitzungsübergreifende Drosselung — verfolgen Sie Missbrauch auf Sitzungsebene persistent, wenn Sie echten Schutz benötigen.

Einrichtung

pnpm install
cp .env.example .env   # fill in TVA_BASE_URL and content-API keys
pnpm build              # builds mcps/shared first, then every domain server, then the gateway

Führen Sie das Gateway aus (alle Tools in einem Prozess):

pnpm start

Oder führen Sie einen einzelnen Domain-Server direkt aus:

node mcps/flight-mcp/dist/index.js

Tool-Referenz

Flug (tva-flight-mcp)

SEARCH_FLIGHTS, GET_POPULAR_FLIGHT_ROUTES, GET_FLIGHT_PRICE_ANALYSIS, GET_MOST_TRAVELED_DESTINATIONS, GET_MOST_BOOKED_DESTINATIONS, GET_BUSIEST_TRAVEL_PERIOD, GET_FLIGHT_SEATMAP_BY_OFFER, GET_FLIGHT_SEATMAP_BY_ORDER, GET_FLIGHT_FARE_UPSELL, PREDICT_FLIGHT_CHOICE, CONFIRM_FLIGHT_PRICE, BOOK_FLIGHT, LIST_MY_FLIGHT_BOOKINGS, GET_FLIGHT_BOOKING, VERIFY_FLIGHT_BOOKING, CANCEL_FLIGHT_BOOKING, GET_FLIGHT_SEAT_PRICE, SELECT_FLIGHT_SEAT, RESEND_FLIGHT_TICKET, GET_FLIGHT_TICKET_DOWNLOAD_LINK, CONFIRM_FLIGHT_PAYMENT

Hotel (tva-hotel-mcp)

SEARCH_HOTELS, SEARCH_HOTEL_CITIES, CHECK_HOTEL_AVAILABILITY, BOOK_HOTEL, LIST_MY_HOTEL_BOOKINGS, GET_HOTEL_BOOKING, CONFIRM_HOTEL_PAYMENT, CANCEL_HOTEL_BOOKING

Visum (tva-visa-mcp)

GET_VISA_INFO_BY_COUNTRY, LIST_MY_VISA_BOOKINGS, GET_VISA_BOOKING, GET_VISA_BOOKING_BY_APPLICATION_UUID, GET_VISA_UPDATES, GET_VISA_UPDATES_BY_BOOKING, GET_VISA_FOLLOWUP_QUESTIONS, ANSWER_VISA_FOLLOWUP, GET_VISA_NOTIFICATIONS, MARK_VISA_NOTIFICATION_READ, GET_VISA_TRANSACTIONS, GET_VISA_TRANSACTION_DETAIL

Einwanderung (tva-immigration-mcp)

GET_IMMIGRATION_INFO_BY_COUNTRY, LIST_MY_IMMIGRATION_BOOKINGS, GET_IMMIGRATION_BOOKING, GET_IMMIGRATION_UPDATES, GET_IMMIGRATION_UPDATES_BY_BOOKING, GET_IMMIGRATION_NOTIFICATIONS, MARK_IMMIGRATION_NOTIFICATION_READ, GET_IMMIGRATION_TRANSACTIONS, GET_IMMIGRATION_TRANSACTION_DETAIL

Konto (tva-account-mcp)

REGISTER_USER, LOGIN_USER, SOCIAL_AUTH_LOGIN, FORGOT_PASSWORD, RESEND_OTP, VERIFY_OTP, LOGOUT_USER, GET_CURRENT_USER, GET_USER_PROFILE, EDIT_USER_PROFILE, CHANGE_PASSWORD, CHECK_DISCOUNT_PROMO, GET_USER_NOTIFICATIONS, CLOSE_ACCOUNT

Verwendet von

MegaMind's NomadSage Reiseagent startet diese Server (über stdio, einen pro Domain) zusammen mit seinen Tools zur Wissensdatenbank-Suche und social-mcp-Antwort — siehe travel-agent/src/lib/tva-client.ts und travel-agent/src/lib/tva-session.ts in diesem Repository.

Available Tools

2 tools
GET_IMMIGRATION_INFO_BY_COUNTRYC

Get immigration information for a specific country.

NameRequiredDescriptionDefault
countryCodeYes

TDQS

C2.8/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden. It states 'Get' which implies a read-only operation, but does not disclose any behavioral traits such as rate limits, authentication needs, data freshness, or error handling. This is a significant gap for a tool with no annotation coverage.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, clear sentence with no wasted words. It is appropriately sized and front-loaded, efficiently conveying the core purpose without unnecessary elaboration.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the complexity of immigration information, lack of annotations, no output schema, and low schema coverage, the description is incomplete. It does not address what type of information is returned, potential limitations, or how it differs from the sibling tool, leaving the agent with insufficient context.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters2/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description coverage is 0%, and the description does not add any meaning beyond the schema. It mentions 'specific country' but does not explain the 'countryCode' parameter's format, valid values, or semantics, failing to compensate for the low schema coverage.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the verb 'Get' and the resource 'immigration information for a specific country', making the purpose understandable. However, it does not explicitly differentiate from its sibling tool GET_VISA_INFO_BY_COUNTRY, which likely provides overlapping or related information, so it misses the highest score.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus its sibling GET_VISA_INFO_BY_COUNTRY, nor does it mention any prerequisites or alternative contexts. It lacks explicit usage instructions, leaving the agent to infer based on tool names alone.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

GET_VISA_INFO_BY_COUNTRYC

Get visa information for a specific country.

NameRequiredDescriptionDefault
countryCodeYes
currencyCodeNo

TDQS

C2.8/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure but only states the basic action without details on permissions, rate limits, error handling, or response format. It doesn't add meaningful context beyond the minimal purpose, failing to compensate for the lack of structured annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, clear sentence with no wasted words, making it highly concise and front-loaded. It efficiently communicates the core purpose without unnecessary elaboration, earning full marks for brevity and structure.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (2 parameters, no output schema, and no annotations), the description is insufficient. It lacks details on parameter usage, behavioral traits, and output expectations, making it incomplete for effective agent operation despite the concise structure.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters2/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description coverage is 0%, and the description doesn't mention any parameters or their meanings. It fails to explain what 'countryCode' and 'currencyCode' represent, their expected formats, or how they influence the output, leaving both parameters undocumented and adding no value beyond the bare schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose with a specific verb ('Get') and resource ('visa information for a specific country'), making it immediately understandable. However, it doesn't explicitly differentiate from its sibling tool 'GET_IMMIGRATION_INFO_BY_COUNTRY', which likely covers related but distinct information, preventing a perfect score.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus its sibling 'GET_IMMIGRATION_INFO_BY_COUNTRY', nor does it mention any prerequisites, alternatives, or exclusions. It only states what the tool does, leaving the agent to infer usage context without explicit direction.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 2 tool updates
    • First observedGET_IMMIGRATION_INFO_BY_COUNTRY
    • First observedGET_VISA_INFO_BY_COUNTRY
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

C2.8/5.0

Scored across 2 tools

Disambiguation2/5

The two tools have unclear boundaries and significant overlap. 'Immigration information' and 'visa information' are closely related concepts in travel, and their descriptions do not clearly differentiate what each tool provides, leading to potential confusion for an agent trying to select the right one.

Naming Consistency5/5

The tool names follow a consistent pattern of uppercase snake_case with a clear 'GET_<domain>_BY_COUNTRY' structure. Both tools use the same verb ('GET') and format, making them predictable and easy to parse.

Tool Count2/5

With only 2 tools, the server feels thin for a 'Travel Agent' domain, which typically involves broader functionality like booking flights, hotels, or checking travel advisories. The limited scope suggests an incomplete or overly narrow implementation.

Completeness2/5

The server is severely incomplete for a travel agent purpose. It lacks core travel operations such as searching for flights, booking accommodations, checking weather, or providing general travel tips, leaving obvious gaps that will hinder agent workflows.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

Appeared in Searches