Skip to main content

Model Context Protocol Server

My travel guides, programmatically.

Every guide on travelerjyoti.com is also available over the Model Context Protocol — a single endpoint AI assistants (Claude Desktop, Claude Code, Cursor, custom GPTs) can call to search, browse, and read first-hand travel content from India to Europe and Asia. Public, read-only, free, no signup.

Endpoint

https://mcp.travelerjyoti.com/mcp
Connect in 60 seconds → Browse all 13 guides

Public read-only <5ms cold start Cloudflare edge (270+ POPs) Protocol 2025-06-18

What is this, in one paragraph?

MCP is a JSON-RPC 2.0 protocol Anthropic shipped in late 2024 and now adopted by OpenAI, Google, Cursor, Continue, and most major AI clients. It lets an AI assistant call your data through structured tools instead of guessing at your URLs.

When a user asks Claude "find me a Switzerland guide on the Swiss Travel Pass", the client looks at its connected MCP servers, sees mine offers a search_guides tool, calls it, and feeds the JSON back to the model. The model answers grounded in my actual content — with a citation back to travelerjyoti.com.

That last bit is the point: every response includes a canonical_url so the AI assistant cites the source. Drives attribution. Drives traffic. Solves the "AI summarised my content without crediting me" problem.

The three tools

Built for travel queries

Each tool returns JSON. Each result includes a canonical URL. Each tool's description is hand-written so the model picks the right one without you nudging it.

list_guides

The index of all 13 guides with metadata — title, country, duration, budget, summary, image URL, dates, canonical URL. The first call most clients make.

Parameters

  • country?: string — optional filter
See example response 
{
  "count": 13,
  "guides": [
    {
      "slug": "swiss-travel-pass-whats-free",
      "title": "What's 100% Free with the Swiss Travel Pass",
      "country": "Switzerland",
      "duration": "3 hours",
      "budget": "Free with pass",
      "canonical_url": "https://travelerjyoti.com/destinations/swiss-travel-pass-whats-free/",
      "updatedAt": "2026-05-19"
    },
    …
  ]
}
get_guide

The full clean Markdown of one guide plus its FAQs and full frontmatter. Use when you already know the slug, or after search_guides identified the best match.

Parameters

  • slug: string — required
See example response 
{
  "slug": "mount-pilatus-free-cap",
  "title": "How to Get a Free Pilatus Cap…",
  "country": "Switzerland",
  "canonical_url": "https://travelerjyoti.com/destinations/mount-pilatus-free-cap/",
  "body": "# Mount Pilatus Free Cap\n\nThis is one of those…",
  "faqs": [
    { "question": "…", "answer": "…" }
  ]
}
search_guides

Keyword-scored search across title, body, and FAQs. Returns ranked matches with a snippet showing why each scored — gives the model enough to decide which to read fully.

Parameters

  • query: string — required
  • limit?: number — min 1, default 5
See example response 
{
  "query": "Swiss Travel Pass",
  "matches": 3,
  "results": [
    {
      "slug": "swiss-travel-pass-whats-free",
      "score": 57,
      "snippet": "…every public train, bus, tram…",
      "canonical_url": "https://travelerjyoti.com/destinations/swiss-travel-pass-whats-free/"
    },
    …
  ]
}

Quick connect

Connect in 60 seconds

Pick your client. Paste. Done.

Claude Code

Local to one project 
claude mcp add travelerjyoti --transport http https://mcp.travelerjyoti.com/mcp

Claude Desktop

Add to claude_desktop_config.json 
{
  "mcpServers": {
    "travelerjyoti": {
      "url": "https://mcp.travelerjyoti.com/mcp"
    }
  }
}

Cursor

Settings → MCP Servers 
{
  "travelerjyoti": {
    "type": "streamable-http",
    "url": "https://mcp.travelerjyoti.com/mcp"
  }
}

curl (raw protocol)

For debugging 
curl -X POST https://mcp.travelerjyoti.com/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call",
       "params":{"name":"search_guides",
                 "arguments":{"query":"Swiss Travel Pass"}}}'

For other clients (Continue, custom GPTs, LangChain, etc.) — point them at the same endpoint with Streamable HTTP transport. Auto-discovery metadata at /.well-known/mcp.json.

FAQ

Common questions

Is it free? Do I need an API key?
Yes, free. No key, no signup, no quota. Public read-only access. I cover the Cloudflare costs because the MCP server has zero data-egress fees inside Cloudflare's network — guides are bundled into the Worker so there's no upstream service to pay for.
How fresh is the data?
The guides are bundled into the Worker at deploy time. Whenever I publish a new guide on travelerjyoti.com, a single npm run deploy redeploys the MCP server with the new content. Latency from publish to MCP availability is usually under 5 minutes.
What's the rate limit?
Cloudflare's free-tier Workers cap is 100,000 requests/day. I'll publish a real limit if usage ever gets close — for now, no enforced limit. If you're planning heavy use (more than a few thousand calls a day), email me and we'll talk.
Can I use the content from this MCP in my product?
Short quotes with attribution + a link back to the canonical URL are welcome. Full-page reproduction is not permitted (see Terms). The MCP server response itself includes a canonical_url in every result specifically so attribution is one field away.
How does this differ from /llms-full.txt?
/llms-full.txt is a one-shot fetch — any AI engine ingests the entire archive at once, no integration needed. MCP is interactive — clients call specific tools with specific arguments and get structured JSON back. Both exist because they serve different audiences: llms-full is for search/answer engines that crawl, MCP is for assistants that converse.
Why isn't this on the official MCP servers directory yet?
Submitted. mcpservers.org takes a few days to review. Once listed it'll show up under 'Travel' (a near-empty category right now).

Built it for myself, made it public.

If you build something with this server, I'd genuinely love to hear about it. Email me, tag me, or just drop a comment on the YouTube channel.

Say hello Discovery JSON About Jyoti