2026 write-up
AI tools for vacation rentals in 2026: where AI fits your stack, and what to build yourself
By Andrew Kitchell, CEO & Founder
AI now sits in four places in a vacation rental stack: inside the PMS, on top of the PMS, on top of your pricing tool, and in your own AI assistant, which reaches your tools through their MCP servers (the standard way an assistant plugs into software). In most of them, the AI works with the tools you already run: it reads them and asks for your yes. In 2023 we doubted one vendor could be the best at every job. Now you don’t need one to be. So pick each tool by how open it is, and build only the part that’s yours.
On this page
What changed since our 2023 write-up?
In January 2023 we published a market map and write-up of the software behind short-term and vacation rentals. We counted at least 19 categories and more than 170 teams, and we named three drivers. Here is where each one stands.
- Scale. Bigger operators need more software. In 2023 that meant more categories of tools. In 2026 it also means AI agents that work across those tools.
- Low barriers to entry. We wrote that an operator can test new technology on a single listing with ease, so a new software team could find 10 or more paying customers relatively easily. The barrier is lower again: now operators build the tools themselves.
- Specialized tools. We wrote that it was unproven that a single provider could deliver the best software across many jobs. APIs and AI make specialized tools the default: the tools now open their data to each other and to AI assistants.
What changed most is how tools connect. In 2023 we wrote that connectivity software teams saw connectivity as the foundation of a new domain of technology. In 2026 that foundation has a standard: the Model Context Protocol (MCP), which Anthropic open-sourced on November 25, 2024. An MCP server lets an AI assistant such as Claude read from, and act in, the software where your data lives.
That will change everyone’s work in this business: the manager who asks an assistant for this week’s arrivals, the revenue manager who approves five price moves before breakfast, and the owner who reads a forecast an assistant put together.
Where do AI tools fit in a vacation rental stack?
Your PMS is still the core of your business, as we put it in 2023. It holds the reservations, listings, owners and channel connections. AI tools now sit at four levels around it.
Your own AI assistant
Reads across your tools through their MCP servers
Examples: Claude, ChatGPT
On top of your pricing tool
Works on the rates your pricing tool sets, and most ask you to approve changes
Examples: IntelliHost, Prometheus iQ, RevSwipr, Sentinel
On top of your PMS
Answers guests, handles reviews and runs team tasks
Examples: Besty AI, HostBuddy AI, Trellis
Inside the PMS
A PMS built around AI
Examples: Boom, PropertyStack, SympleHost
Your PMS
Holds your reservations, listings, owners and channel connections
- Inside the PMS. Some newer PMSs are built around AI from the start. Boom calls itself the AI vacation rental platform, PropertyStack sells agentic property management, and SympleHost calls itself the AI co-host for short-term rentals and is a full PMS too.
- On top of your PMS. These tools connect to what you already run and take on the guest side: messages, reviews and team tasks. Besty AI says it plainly: keep your existing PMS. HostBuddy AI and Trellis work at this level too.
Which vacation rental tools have an MCP server?
Here are tools with an official one, live or announced, as of November 18, 2026.
| Tool | MCP server |
|---|---|
| WheelhousePricing and revenue management | Out of beta and free for now. Does everything the app and API do, with a read-only switch |
| HospitablePMS | Live since April 2026 |
| GuestyPMS | Beta, every tool read-only |
| RentalReadyPMS | Works with Claude, OpenAI and Gemini |
| LodgifyVacation rental software | Coming soon |
| PriceLabsPricing | Shipped July 2026 |
| BeyondPricing | Beta, sign up to join |
| IntelliHostAI pricing and listings | On the Agent plan |
| AirDNAMarket data (Adapt) | Beta |
| ConduitGuest messaging | On the v3 plan |
| TrellisGuest messaging and operations | Live for Claude, Cursor and Codex |
| AutohostGuest screening | Live as a ChatGPT app and Claude connector |
| SendSquaredGuest CRM and marketing | On enterprise plans, read-only by default |
| JetstreamChannel manager | Live |
| NowistayAI co-host and channel manager | Live as a ChatGPT app and Claude connector |
| StayFiGuest WiFi and direct bookings | Live |
Hospitable calls its server the first official MCP server from a short-term rental platform, and it comes with every Hospitable plan except Essentials. Nowistay’s works on its full plan. Hostaway has a public API that creates and updates listings and reservations and reads calendars, so your developer can build on it today.
What do the new revenue management tools do?
The newest revenue management tools work with the pricing tool you already run, and they ask you to approve the next move.
- Sentinel sits on top of pricing tools such as PriceLabs and Wheelhouse, and on your PMS. Its rule: nothing changes without a human’s yes.
- Prometheus iQ works with PriceLabs or Wheelhouse and has its suggestions ready for your approval each morning.
- RevSwipr shows up to five evidence-backed moves a day, and you approve each one with a swipe. It started as Swipr, a project at our July 2026 hackathon, and calls itself the hackathon’s winner. RevSwipr is an independent product.
- IntelliHost works on top of PriceLabs and Wheelhouse if you keep them, and, on its Agent plan, with Claude, ChatGPT and other MCP tools.
So you keep the pricing engine you have. The new layer explains it, spots what the engine missed, and puts the decision in front of you.
Is the new PMS just Claude connected to specialty software?
No. An assistant connected to your tools is a great way to work across them, but it needs a record to read from. Reservations, owner statements, channel connections and payouts have to stay right when the assistant is switched off, and the PMS is where they stay right.
What shrinks is the case for buying every feature from one vendor. When your PMS and your specialty tools all open their data, an assistant can pull from the best tool for each job.
Can operators build their own tools now?
Yes, for the work only your team does. At our hackathon, July 20 to 24, 2026, teams built tools on Wheelhouse’s API in five days. A few of them:
- Revenue Central, by Alisa Kück and Curated Property: owner reports, forecasts and pricing settings in one place, with AI pushing changes straight into Wheelhouse.
- Project Reins, by Jeff Brown and IntelliHost: watches listings against Wheelhouse comp-set data and lets you approve real pricing decisions with one tap.
- Gap Rescue Radar, by Deng Aguer Bul: finds unbooked gaps in the next 180 days, works out whether the cause is minimum-stay rules, booking patterns, lead time or price, and recommends a fix you control instead of a discount by default.
- Helm, by Mannas Narang: two AI agents on live Wheelhouse data. One fixes or flags underperforming units with human approval. The other drafts demand campaigns for units that are priced right but still aren’t booking.
What are the tradeoffs of building your own tools?
Building has never been cheaper, but it still costs you in three places.
- Upkeep. When a vendor changes its API, the fix is yours. A tool your team built is a tool your team maintains.
- Control. An assistant with write access can act on a wrong reading: change a price, message a guest, block a date. So start read-only, and when you let it make changes, keep a person approving each one.
- The hard parts. A pricing model and market data take years to get right (our engine has 8+ years of research behind it). Our view: use a vendor’s for those, and build the parts that are only yours, like your owner reports, your team’s checklists, or how you want gaps handled.
So before you pick any tool in 2026, ask three questions. Does my plan include its API? Does it have an MCP server? Can an assistant only read, or can it make changes too?
Frequently asked questions
How is Airbnb using AI?
Airbnb runs an AI assistant for customer support, now in more than 50 languages. In Q3 2026, about 40% of all its customer support issues were resolved without a human agent, up from a third in Q2 2026.
Build on a pricing engine you don’t have to maintain
The Revenue Management API comes with every Wheelhouse plan, including the free one, and pulls and pushes data for paid listings. More than 300 customers use it already. Our MCP server is out of beta and free for now. It does everything the app and the API do, and one switch makes it read-only. Connect it to Claude in under two minutes, or start with the API docs.