Hospitable's free Essentials plan gives you AI drafts capped at 10 uses a week, the Host plan adds the Message AI Assistant and MCP access for AI agents, and Inbox AI, which can send replies by itself, is listed for the top Mogul plan only. Paid prices are not printed on the pricing page.
That split matters because "Hospitable has AI" describes three different things: a drafting button you press, an assistant that works inside your inbox, and an auto-reply engine that answers guests while you sleep. Most small hosts need the first, a few need the second, and the third carries risks you should test before switching it on. What follows separates them, using Hospitable's own pricing and help pages as read in October 2026.
The four plans and where the AI starts
Hospitable sells four tiers: Essentials (free), Host, Professional and Mogul. Its pricing page describes Host as the plan for a solo host, Professional as the one for scaling a portfolio and coordinating a team, and Mogul as the top tier with "top-tier AI, smart devices, and full owner management tools". There is a 14-day free trial that needs no card, and an annual billing toggle marked "Save 12%".
When I loaded the page, the Host, Professional and Mogul prices showed as blank placeholders, so I won't quote a figure. Third-party review sites do print numbers, but they disagree with each other and go stale, so treat any dollar amount you read outside hospitable.com as a lead to check, not a fact.
| Plan | AI features Hospitable names | What it means in practice |
|---|---|---|
| Essentials (free) | Suggest with AI and Improve with AI, capped at 10 combined uses a week; AI chat for support | Enough to try drafting, not enough to run a busy inbox on |
| Host | Message AI Assistant, Dynamic Pricing, running your business through AI agents with MCP | Where drafting stops being rationed |
| Professional | Everything in Host, plus team and portfolio tools; no extra AI feature named in its summary | Pay for it for the team features, not for the AI |
| Mogul | Everything in Professional, plus Inbox AI Messaging Assistant (including auto reply) | The only tier where AI can answer a guest without you |
One caution on the table. The pricing page also has a row for AI Copilot, but when I read the page the tick marks did not come through, so I can't tell you which plans include it; read that row on the live comparison table yourself. The Knowledge Hub is easier: Hospitable's Essentials plan overview lists it as included on the free plan and describes it as what powers Inbox AI, so you can build it before paying for anything. The same overview lists AI auto-replies, unlimited Suggest with AI and Hospitable's MCP among the features Essentials does not get.
What the free Essentials cap really limits
Hospitable's help article on AI-suggested replies describes two buttons in the Inbox. Suggest with AI sits in the compose area and drafts a reply to the guest's last message in a few seconds. Improve with AI rewrites text you have already typed, including a rough bullet-point note to yourself. If you dislike the result, Undo restores your original wording.
According to the help centre, Essentials users get 10 uses a week across both buttons combined. One review I read claims the cap covers only suggestions; Hospitable's wording says combined, and I'd go with the vendor. The practical effect is easy to work out. A host with two listings might get six or seven guest conversations a week, and each conversation often needs two or three replies. A cap of 10 runs out by Tuesday.
A quick sum for a two-listing host
Say a host has two listings and 5 stays arriving in a week. Each stay produces roughly one pre-arrival question, one check-in reply, one mid-stay question and a checkout thank-you. That is about 4 replies a stay, 20 replies a week. If the host uses Suggest with AI only on the awkward half, that is 10 uses, which fits the cap exactly. Add one Improve with AI polish on a complaint and the free plan is already over. These numbers are illustrative; count your own last four weeks of messages before you decide.
What the AI reads before it drafts
The help article says Suggest with AI draws on five things: the listing, the reservation, the guest, your Knowledge Hub, and how you have answered similar messages before. That list is the whole quality story. A draft is only as accurate as the Knowledge Hub entries behind it.
So before you judge the AI, fill in the Knowledge Hub honestly: Wi-Fi name, parking rules, bin days, heating quirks, the key-safe arrangements, and the three things guests always ask. Then test with real questions. Here is the kind of exchange I would run on day one of a trial. The sample below is illustrative, not captured output.
Guest message: "Hi, we land at 11am. Is there any chance we can drop bags
before the 3pm check-in? Also is there parking for a van?"
Reasonable AI draft: "Hi [first name], thanks for letting us know. You're
welcome to leave bags from 11am, but the flat won't be ready until 3pm, so
the key safe details will arrive at check-in time. There is room for one van
on the driveway; please don't block the shared gate. Safe travels!"
What to fix: the draft promised bag drop but your Knowledge Hub says the
cleaner is in until 2pm. Edit the answer, then add a Knowledge Hub line
about early bag drop so the next draft is right.If the draft invents a policy that your Knowledge Hub does not contain, that is the signal to stop and fix the source material, not to upgrade the plan. The house manual tutorial shows how to turn your own notes into the kind of content the Knowledge Hub needs.
Improve with AI: a before and after
Improve with AI is the underrated button, because it starts from your facts rather than the tool's guess. You jot the points, it makes them polite. An illustrative example, written as the host would type it and as a tidy reply might come back:
Before (typed in a hurry):
"cleaner late, flat ready 4 not 3, sorry, leave bags in hall, wifi on fridge"
After (illustrative output):
"Hi [first name], a quick update: our cleaner is running a little behind, so
the flat will be ready at 4pm rather than 3pm. Sorry for the change. You're
welcome to leave your bags in the hall, and the Wi-Fi details are on the
fridge. See you soon!"Check one thing in the result: it kept every fact you typed and added none. If it adds "a welcome basket" you never mentioned, undo it. Because each use counts towards the free weekly cap, a host on Essentials may prefer to spend those 10 uses on this kind of fact-led rewrite than on long cold drafts.
Message AI Assistant on the Host plan
Hospitable lists the Message AI Assistant with the Host plan, alongside Dynamic Pricing. Its help pages do not give a separate weekly limit for it that I could find, and the pages I read do not spell out every detail of how it differs from the two Essentials buttons. The safest reading is that Host is where drafting is no longer rationed, and that the assistant works as a drafting aid inside the inbox, with you approving each message.
The test that decides whether Host is worth it is simple: count how many replies you wrote by hand last month, and how many of those a good draft could have covered with a light edit. If it is more than about 40 a month, the time saving is likely real. If it is 10 or fewer, the free plan's cap may be enough, and your money is better spent elsewhere.
Dynamic Pricing comes with the same plan
Host also includes Dynamic Pricing. The pricing table says it is free for 3 bookings in the first 30 days, then billed per property. That is separate from the AI messaging, so don't let it inflate the cost of "the AI". If you already use another pricing tool, work out whether the plan's price still makes sense without it. The comparison of Lodgify, Hospitable, Guesty and HostBuddy covers how the alternatives bundle these.
Inbox AI and auto reply on Mogul
This is the feature people mean when they ask whether AI can "run my messages". The help centre states that Inbox AI is available to Mogul customers only, and that AI Auto Reply is part of it: it sends replies to guest messages automatically, based on your settings, the property's context and your past conversations. The article says Auto Reply only answers when all its conditions are met, will not respond to certain sensitive topics, and records the reason in the conversation notes (with an email to you) when it declines. You can also pause it for a single reservation.
Two details deserve a close read. The first is scope: you can create several settings groups and scope each one to specific properties, booking stages or time ranges, and only the first applicable group runs on a conversation. The second is timing: a Timing tab sets when Auto Reply is active. A sensible starting configuration is narrow. Let it answer only practical questions such as Wi-Fi, parking and check-in instructions, only outside your own waking hours, and leave anything about money, damage, safety or complaints for you.
A staged way to switch it on
- Week one, drafts only. Use suggested replies and note how often you change them. Count edits, not impressions.
- Week two, narrow auto reply. Allow one topic, such as arrival logistics, for one listing. Read every sent reply the next morning.
- Week three, widen. Add the next topic only if no reply needed an apology or a correction.
- Always, keep the exits. Know how to switch it off in two clicks, and know what a guest sees if it errs.
Hosts often ask whether automated replies break Airbnb's rules. I did not find a page in Hospitable's help centre that covers Airbnb policy on automated replies, so check Airbnb's current guidance on third-party tools yourself rather than assuming. The tutorial on testing the AI before you switch automated Airbnb messages on gives a fuller test plan.
MCP access: for hosts who already live in an AI assistant
The Host plan description says you can "run your business through AI agents with MCP". MCP (Model Context Protocol) is an open standard that lets an AI assistant connect to an outside tool and read or change its data. In plain terms, you could ask an assistant such as Claude or ChatGPT, "which guests arrive this weekend and have not had a check-in message?" and have it look inside Hospitable for the answer.
It is a feature for the minority. If you do not already use an AI assistant for business tasks every week, you will not feel its absence. If you do, be careful with write access: an assistant that can change reservations or send messages through MCP needs the same narrow permissions and morning review as auto reply. Hospitable's pricing page names the feature but I did not find a help article that lists exactly what an agent can and cannot do, so test with read-only questions first.
Choosing a plan from your own numbers
A three-question rule decides most cases.
- How many guest messages a month? Under about 40 and the free cap, plus your own saved replies, may do. Over that, uncapped drafting is likely to pay for itself in time.
- Do you answer at night? If replies at 2am are costing you sleep or bookings, only auto reply solves that, and that means Mogul.
- Do you manage other people's properties? Then the Professional features matter more than the AI.
Three hosts, three answers
A host with one flat and a handful of stays a month: stay on Essentials, keep a good set of saved replies, and use the 10 weekly AI uses on the awkward messages. A host with four cottages and steady summer bookings: Host is the likely fit, because capped drafting would bite in July. A co-host running a dozen owners' properties, with cleaners to coordinate and guests arriving at all hours: the Mogul tier is where auto reply and owner tools sit, provided the narrow staged rollout above goes cleanly. The co-host business tutorial covers the multi-property side.
A fourth case: a host who gets mostly the same five questions. Before paying for any AI plan, write those five answers as saved templates; the nine Airbnb message templates are a starting point. Templates cost nothing and never invent an answer.
A 14-day trial plan that tells you the truth
Hospitable offers a 14-day trial with no card. Fourteen days is short, so go in with a script.
- Days 1 to 2: fill the Knowledge Hub for your busiest listing. Time yourself; if it takes over three hours, that is a real cost of the tool.
- Days 3 to 7: run Suggest with AI on every guest message. Log each draft as sent as is, edited lightly, edited heavily, or discarded.
- Days 8 to 10: feed it your five hardest historical messages: a refund demand, a noisy-neighbour complaint, a midnight late arrival, a damaged item, a review threat. Do not send these; just read the drafts.
- Days 11 to 13: if your trial includes Inbox AI, run it for one topic only.
- Day 14: total the log. If more than a third of drafts needed heavy edits, the problem is your Knowledge Hub or the tool, and neither is fixed by paying more.
A scoring sheet you can copy
Messages drafted: ___
Sent as is: ___ (target: 40% or more)
Light edit: ___
Heavy edit or discarded: ___ (red flag above 33%)
Invented a fact/policy: ___ (red flag at any)
Minutes saved per msg: ___ (your estimate)
Plan price per month: $___
Break-even: price / (minutes saved x messages x your hourly value)Fill it in with real numbers. Say the log shows 60 drafts: 24 sent as is, 21 lightly edited, 12 heavily edited, 3 discarded, and none inventing a policy. That is 40 per cent untouched and 25 per cent heavy or discarded, a passing grade. If those 60 messages saved about 3 minutes each, that is 180 minutes, three hours. Whether three hours a month justifies a plan depends on what your hour is worth and what the plan costs; that is why the sheet ends with a break-even line you complete yourself.
Where the AI goes wrong in a holiday let
Four failures recur with message AI in short lets. It states a policy you do not have, such as a flexible cancellation. It answers a safety question casually, such as how the gas hob works. It misses emotion, replying cheerfully to a message that is really a complaint. And it repeats information the guest has already been given, which feels automated. All four are caught by the same habit: for the first month, read what it drafts or sent.
A realistic example of the third failure: a guest writes, "The shower is still cold after 20 minutes, we're a bit disappointed." A drafting tool trained on cheerful past replies may open with "Great question!" The fix is a line in your Knowledge Hub or instructions that any message mentioning a fault gets an apology first and a concrete next step second, and in the meantime you edit the draft by hand.
The safety line is the one I would never relax. Anything involving fire, gas, locks, medical needs or damage claims should land in your own hands. Pair that with the calm replies tutorial for complaints so you have a good human answer ready.
How to tell the trial worked after you cancel or keep
A fortnight of drafts feels impressive and proves little. Judge the result a month later instead. Open your sent messages from the trial period and compare response times with the month before: did guests get answers faster, and did the number of follow-up questions ("sorry, where exactly is the key safe?") fall? Look at your reviews for the same weeks, and search them for the word "communication", since that is the score messaging tools exist to protect. If response time improved and complaints did not rise, the plan is doing its job. If you spent more time correcting than you saved writing, go back to saved replies and reconsider in a season when volume justifies it.
What to check on the live pages before you pay
- The current price of the plan you want, monthly and annual, on hospitable.com.
- Which plan ticks AI Copilot in the comparison table.
- Whether the weekly cap on the free plan is still 10 combined uses.
- Whether Inbox AI is still Mogul only. Vendors move features between tiers.
- Airbnb's current stance on automated replies from approved software.
- How to cancel and whether the plan renews automatically.
If you would rather map your own message volume and decide with someone, that is what my AI implementation consultation covers. Otherwise, the free trial and the scoring sheet above will get you most of the way.
Hospitable AI questions hosts ask before upgrading
Is Hospitable's AI free?
Partly. The free Essentials plan lets you use Suggest with AI and Improve with AI, but Hospitable's help centre caps the two combined at 10 uses a week. Paid plans lift that cap, and the Message AI Assistant and Inbox AI sit on higher plans. Confirm the current limits in your own account.
Can Hospitable's AI send replies without me approving them?
Yes, on the top plan. Inbox AI includes an auto reply feature that sends answers on its own, based on your settings, property context and past conversations, and the help centre lists it for Mogul only. Suggested replies on lower plans stay as drafts that you review and send yourself.
Does Hospitable list its prices?
Not clearly. When I loaded the pricing page for this tutorial, the paid plans showed a blank price placeholder and only Essentials showed as free. A monthly or annual toggle exists, with annual marked as saving 12 per cent. Read the live page, or start the 14-day trial, to see your real figure.
What is MCP in Hospitable and do I need it?
MCP is a standard that lets an AI assistant connect to a tool and act on its data. Hospitable lists running your business through AI agents with MCP on its Host plan. A host who only wants faster guest replies can ignore it; it suits people who already work in an AI assistant daily.
Further reads
- Can Airbnb Hosts Use AI to Reply to Guests? What Airbnb Allows — What Airbnb's own rules say about AI-written guest replies.
- Lodgify AI Co-Host: What It Does and What to Test First — The same plan-by-plan test applied to Lodgify's AI.
- Airbnb No-Show Wants a Refund: Reply Scripts Drafted With AI — Refund requests to keep out of auto reply, with scripts.
- How B&B Owners Can Automate Airbnb and Booking.com Messages — Automating Airbnb and Booking.com messages without losing the personal touch.
- Airbnb Cleaning Schedule: Coordinate Cleaners and Changeovers — Another job to automate once guest messages are settled.
- Is HostBuddy AI Worth It for an Airbnb Host? — An independent test of a messaging add-on that sits on top of a PMS.
- Best Accounting Software for Airbnb Hosts: Fees, Payouts and AI Receipts — Compare Xero, QuickBooks, Wave and Zoho Books on host payouts, platform fees and how well their AI reads receipts and picks categories.
- Holiday Let Website: Get Direct Bookings and Repeat Guests With AI — A page-by-page plan for a direct-booking holiday let site, with Lodgify and Wix compared, a calendar-sync test and the sum for break-even.
- Vacation Rental Analytics: Occupancy and Rate Reports With AI — Export bookings, split stays across months, work out occupancy and nightly rate in a sheet, then have AI write the monthly summary.
- AI Agents for Short-Term Rentals: What a Host Can Do Today — Ten host jobs rated from hand over to keep, with the checks that stop an AI agent promising refunds or double-booking your cleaner.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Hospitable pricing page and help centre articles on AI-suggested replies and AI auto replies, read in October 2026.