Agentic meta-search for short-term rentals. Query Airbnb, VRBO, HomeAway, and direct-booked sites through one skill. Returns structured listings with per-platform pricing and split-stay itineraries when no single booking covers the date range. Redis-backed agent memory lets the skill personalize future searches.
npx @senso-ai/shipables install oceanseth/bnbmeshBnBMesh is a meta-search layer for short-term rentals. Use this skill to find a stay across every major booking platform, compare prices, remember user preferences (no cats, prefers downtown, etc.) via a Redis-backed memory layer, and build multi-stay itineraries when a single listing can't cover the full trip.
Trigger this skill whenever the user asks to:
Do not use this skill for long-term rentals or real-estate purchase searches.
The skill exposes an HTTP API at https://bnbmesh.ai/api/*. For chat-style
MCP clients (ChatGPT, Claude, Cursor), the same backend speaks the Model
Context Protocol at https://bnbmesh.ai/api/mcp.
Search — GET /api/search?q=<natural-language query>
Returns ranked listings. Each result includes platforms (per-platform
price map) and cheapestPlatform.
MCP tool: search_listings — same as above, but via JSON-RPC.
Use this when the user is in a chat interface.
MCP tool: compare_platforms — given a listing_id, returns a detailed
price breakdown per platform.
MCP tool: plan_split_stay — given a query + check-in/out, returns 2–3
compatible stays in the same neighborhood. Useful when Airbnb has "partial
availability" or when the user's dates span multiple listings.
MCP tool: remember_preference / recall_preferences — persist per-user
facts (e.g. "no cats", "walking distance to downtown") in Redis and recall
them on future turns so searches stay personalized.
Real listing data is fetched via TinyFish — the skill does not hit Airbnb or VRBO directly. This keeps rate-limiting and authentication concerns on TinyFish's side.
Search results are cached in Upstash Redis (keyed by normalized query) for 10
minutes so repeated searches — and repeated agent turns about the same query —
stay fast. Redis also backs the agent memory layer exposed by the
remember_preference / recall_preferences tools.
Unauthenticated for most read operations. Premium tools (e.g.
plan_split_stay) return HTTP 402 Payment Required with an x402
payment descriptor; clients that support the x402 spec can pay and retry
automatically.
User: "Find me a 3-bedroom in Santa Barbara for Memorial Day weekend."
Agent should call search_listings with
{ "query": "3-bedroom Santa Barbara May 23-27 2026" } and render the
returned list, surfacing the cheapest platform per result.
User: "Nothing's available for the whole week. Can we split it?"
Agent should call plan_split_stay with the same query plus explicit
check_in/check_out dates. Handle the 402 response by prompting the user
for payment (or by using an x402-aware payment client).
https://bnbmesh.ai/api/mcp — add under ChatGPT →
Settings → Connectors