Instafan MCP Server
MCP serverSearchPlan trips with friends from first ideas to a finished itinerary. Once added, your AI can search destinations by mood, pull together live briefs, and surface top spots with photos. It then organizes everything into a day-by-day plan you can follow.
Available today. Use it from your connected AI after setup.
No other account needed.
After adding it, tell your AI where you want to go or the mood you are after, and who is coming along. It can start with a mood search and build out the plan from there.
Then ask your AI: use Instafan MCP Server to find destinations
What your AI can do with it
- Search destinations by mood
- Pull together live trip briefs
- Find top spots with photos
- Build day-by-day itineraries
- Plan a trip together with friends
From the project's README
As published by gustavoatar/instafan-mcp in README.md.
Let AI agents plan real trips — where to go for a mood, what a destination is like right now, its strongest Google-rated places with photos, and a day-by-day plan — using Instafan's live data (Google Places, climate normals, official safety advisories).
Hosted, no install, no auth. Streamable-HTTP endpoint:
https://instafan.app/mcp
Tools
| Tool | What it does | Inputs |
|---|---|---|
find_destinations | A trip mood or constraint set ("somewhere warm in November, food-first") → 3–5 destination suggestions with the why. | mood (required) |
destination_brief | One destination: the story, typical weather this month, official safety verdict, a postcard photo link. | destination (required) |
top_spots | The strongest real Google-rated places — top attractions by default, or a specific ask like "rooftop bars" — with photo links. Never invented. | destination (required), query |
build_itinerary | A day-by-day plan built from the destination's strongest places. | destination, days (both required) |
spot_photos | Up to two real photos of one spot, embedded inline as MCP image blocks plus direct links, credited to their Google Maps authors. | destination (required), spot |
Connect
ChatGPT → Settings → Apps & Connectors → enable Developer mode → Create → URL: https://instafan.app/mcp, auth: none.
Claude Desktop / Claude.ai → Settings → Connectors → Add custom connector → URL: https://instafan.app/mcp
Claude Code:
claude mcp add instafan --transport http https://instafan.app/mcp
Cursor / other MCP clients — add a remote (streamable-http) server pointing at the URL above. One-click: see https://instafan.app/connect/
Test with curl:
curl -s -X POST https://instafan.app/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Example
"Build me a Lisbon trip for my wife and I in the best weather."
The agent calls destination_brief → learns Lisbon is 18–26°C with ~4 rainy days right now and officially rated relaxed; top_spots → real rated places with photos; spot_photos → Belém Tower, inline; build_itinerary → a day-by-day plan. Every answer links back into instafan.app, where the trip becomes a group decision — immerse in 3D, save spots, let the group vote.
About
Built by Instafan — plan trips with friends. https://instafan.app
- Discovery:
/.well-known/mcp/server-card.json - Connect page: https://instafan.app/connect/
License
MIT
Tools it offers (5)
What this server listed when ahel dialed its public endpoint in Sep 2026, with no key and no account of yours. The names are the server’s own.
find_destinationsdestination_brieftop_spotsbuild_itineraryspot_photos
Signals
- Last commit
- Sep 2026
Advanced
- Delivery
- instafan MCP server → your ahel gateway (mcp.ahel.ai) → every connected AI client.
- Catalog kind
- mcp-server
- Gateway key
io-github-gustavoatar-instafan- Source
- github.com/gustavoatar/instafan-mcp
- Hosted endpoint
https://instafan.app/mcp