Most travel apps want to be your planning tool. HotelTonight wants to be the thing you open in a taxi when the meeting overran and you are not getting home tonight. It has been refusing to broaden that ambition since 2010, and the refusal is why it remains good.
The product, in one paragraph
Open the app. It knows where you are. It shows a short, curated list of hotels with rooms unsold for tonight, at rates below what those hotels are publishing elsewhere. You pick one, pay, and walk in. There is no filter tree, no map of three hundred pins, no sorting by eleven criteria. The whole transaction takes under a minute.
That sounds trivial until you try to do the same thing on a general OTA at 6pm with 12% battery.
Why the rates are lower
A hotel room unsold at 4pm is worth nothing at midnight. HotelTonight exists to convert that zero into something, and hotels will discount steeply to make it happen — provided the discount is confined to a same-day channel where their regular customers will not learn to wait for it.
This is the same logic as an opaque rate, but with one important difference: the properties are named. You can see the hotel, read about it, look at the photographs, and decide. You give up planning time rather than information, which is a far better trade than the one Hotwire and Priceline offer.
The discounts are not uniformly dramatic. In saturated markets on quiet weeknights they can be substantial. On a Friday in a city hosting a conference, there may be no rooms at all, at any price. The model only produces bargains when there is slack.
The coverage problem
HotelTonight works in large cities with deep hotel supply. It does not work in small ones. Open the app in a regional town and you will either see nothing or see two properties twenty kilometres away. This is not a bug — the mathematics of last-minute inventory require a lot of rooms — but it means the app cannot be your default. It is a tool you carry for the cities where it works.
Booking further ahead is possible within a limited window, but the inventory thins the further out you look, and the discount thins with it. Using HotelTonight to book a month out is using the wrong tool.
The Airbnb question
Airbnb acquired HotelTonight in 2019, and in 2026 the integration remains shallow. You cannot meaningfully earn or spend anything across the two, the accounts are largely separate, and Airbnb's own hotel listings do not surface HotelTonight rates. Whatever the strategic logic was, the practical answer for a traveller is to treat them as unrelated products.
There is a plausible reading in which this is deliberate and correct: HotelTonight's value is its narrowness, and folding it into Airbnb's interface would destroy exactly the thing that makes it work.
Who it is for
People whose plans change late. Delayed flights, extended meetings, a night out that went longer than intended.
City-based travellers who move between large metros where the inventory pool is deep.
Anyone who finds general booking sites exhausting. A short curated list is a feature, not a limitation.
It is not for trip planning, not for small towns, not for anywhere you need a specific room type, and not for accumulating hotel status — like other third-party rates, these bookings rarely earn chain credit.
What the discount actually looks like
It is worth putting numbers on a claim the app makes visually rather than numerically.
The mechanism only produces a discount when a hotel has rooms it will not otherwise sell. That means the size of the saving tracks occupancy, and occupancy is predictable: Sunday to Wednesday in a business city, any night in a leisure city out of season, and anywhere with a new hotel trying to build a reputation.
It also means the reverse. Friday and Saturday in a popular city, anywhere hosting a conference or a festival, and the whole of a peak season produce either no inventory or rates no better than the hotel's published price. The app does not hide this — you simply see a short list or an empty one — but people who have had one spectacular result tend to expect it every time.
The practical rule: HotelTonight is a tool for filling a gap in a soft market, not a strategy for cheap hotels.
Booking behaviour worth knowing
Rates are prepaid at booking. Cancellation terms are set per property and are frequently tighter than you would get booking the same hotel on a conventional site, which is the trade for the price.
Check-in also occasionally surprises people. A same-day third-party booking sometimes reaches the property's system minutes before you do, and the front desk may need a moment to find it. Having the confirmation on screen rather than in an email you cannot load on hotel wifi is worth the ten seconds of preparation.
The verdict
6.9 reflects a genuine conflict. On execution — interface, speed, clarity, curation — HotelTonight is one of the best-made products in this entire ranking and would score well into the eights. On coverage and applicability it is one of the narrowest. A tool you can only use in certain cities, on certain days, for a decision you are making within hours, cannot rank alongside platforms that work everywhere.
Install it anyway. The two nights a year it saves you will pay for the icon on your home screen.
