The SwitchBot Contact Sensor is the honest answer: it detects the door opening, detects movement in front of it, and sounds its own alert with nothing else installed. Everything else marketed as hub-free on this page needs something. The table below states exactly what, per product, before you spend anything.
What does “no hub” actually mean?
Three different things, and vendors use the phrase for all three without distinguishing between them. This is the single most common reason a sensor gets bought, opened, and returned.
- Truly standalone. The device does its whole job with no second box. It pairs to your phone over Bluetooth, or it does something local and audible, and it is useful the moment the battery tab comes out. This is rare and it is what most buyers think they are getting.
- No hub to buy, but a border router you must already own. Thread devices need a Thread border router. Apple TVs, HomePods, several Nest and Echo devices and SmartThings hardware are border routers, so if one is already in your house the sensor genuinely joins with no additional purchase. If not, you are buying a hub, whatever the listing calls it.
- No hub because it uses somebody else's network.The Ring Glass Break Sensor rides Amazon Sidewalk. There is no base station, which is true and useful, and there is a third-party mesh carrying your device's traffic, which is also true and is a different proposition entirely.
The category-three case is the one to read twice. “No base station required” and “no infrastructure required” are not the same claim. One means you save a purchase; the other means your alert path depends on a network you neither own nor administer.
What “no hub” means for each of these
Compiled from manufacturer specifications and published protocol documentation, July 2026.| Sensor | What “no hub” actually means for this product | Protocol | Works with nothing else installed? | Works offline? | Power |
|---|
| SwitchBot Contact Sensor | Truly standalone for the audible alert and phone-in-range use. A SwitchBot hub is needed only for notifications while you are out. | Bluetooth | Yes | Yes, the local alert needs no network at all | Replaceable cells |
|---|
| Aqara Door and Window Sensor P2 | No Aqara hub needed, but a Thread border router is mandatory. If you own an Apple TV 4K, HomePod mini, Nest Hub (2nd gen) or SmartThings Station you already have one; if not, you are buying one. | Matter over Thread | No | Yes, local automations | 1x CR2032 |
|---|
| Ring Glass Break Sensor | No Ring Alarm base station needed, because it uses the Amazon Sidewalk network instead. You skip a purchase and take on a third-party network dependency. | Amazon Sidewalk | Yes, given Sidewalk coverage | No, the alert path is not yours | Replaceable cells |
|---|
| Aqara Multi-State Sensor P100 | Not hub-free. It needs a Thread border router or an Aqara Zigbee hub. It is here because it is frequently shopped alongside the P2 and buyers assume it inherits the same requirements. | Thread or Zigbee | No | Yes, once joined | Coin cell |
|---|
| SwitchBot mmWave Radar Motion Sensor | Pairs to your phone directly over Bluetooth, but a SwitchBot hub is required before it can drive any automation. Mains power is also required in practice. | Bluetooth | Partly, for local readings only | Yes, but it cannot tell anyone | Mains |
|---|
Why the SwitchBot wins a page it is not the best sensor on
On raw capability the Aqara P2 is the better device and it wins our main door and window sensor roundup on exactly those grounds. It is second here because this page answers a narrower question: what should you buy if you are not going to install any infrastructure?
Under that constraint the SwitchBot is the only pick that fully clears the bar. It detects the opening, detects movement in front of it, and makes a noise where the door is. On a garage side door, a shed, a rented room or a basement door, an audible local alert is frequently more useful than a phone notification, because the notification arrives somewhere you are not. Its limits are real and stated in the same breath: Bluetooth range is short in a house with thick internal walls, and without a SwitchBot hub it will not reach you at work.
The border router trap
Thread devices are the single biggest source of the confusion this page exists to fix, because the listing is technically accurate and practically misleading. “No hub required” on a Thread sensor means “no proprietary vendor hub required”. It does not mean the device works on its own. Thread is a mesh network, and a mesh needs a router that bridges it to your LAN.
The redeeming detail is that many people already own one without knowing it. An Apple TV 4K, a HomePod mini, a Google Nest Hub (2nd gen), a SmartThings Station and the Aqara Hub M3 and M200 all act as Thread border routers. If any of those is plugged in, a Thread sensor genuinely is a no-extra-purchase device, and it is the better buy. Check before you order, not after. The full mechanism is in our Matter and Thread explainer.
Standalone alarms versus connected sensors
There is a legitimate product category this page will not pretend into existence: the truly standalone door alarm that also notifies your phone from anywhere. Physics is unhelpful here. Reaching your phone from away requires either your home network or somebody else's, and a coin cell cannot power a Wi-Fi radio for a year. Every product that solves this solves it by adding a hub, riding a shared mesh, or having mains power.
So the honest framing is a choice, not a compromise. Either you want a local audible response, which needs nothing and works forever, or you want a remote notification, which needs infrastructure. Deciding which of those you are actually buying, before you shop, removes about ninety percent of the disappointment in this category.
What hub-free costs you, in specifics
Going without infrastructure is a legitimate decision, and it has four concrete consequences worth knowing before rather than after.
Range collapses.Bluetooth is the radio that makes standalone sensors possible, and its practical indoor range is a fraction of Zigbee's or Thread's because there is no mesh to relay through. Two plasterboard walls and a run of ductwork will cost you a connection that worked fine in the same room. A mesh network solves this structurally: every mains-powered node repeats for the battery ones. A standalone sensor has nothing to repeat through, so the only fix is standing closer.
Automation stops at the device.A hub is the thing that holds rules. With no hub there is nowhere for “if the back door opens after 11pm, turn the hall light on” to live. You get an alert and a human decision, which is fine for one door and useless as a system.
State history disappears.Hub-based sensors log every open and close, so you can answer “was the garage door left open yesterday afternoon?” retroactively. A standalone alert is an event, not a record.
Cost per opening goes the wrong way. A standalone sensor costs more per unit than a bare mesh endpoint because it carries its own alerting hardware. On one door that is irrelevant. On nine, the hub you avoided buying would have paid for itself twice over, which is the arithmetic worked through in the main roundup.
None of that argues against buying one. It argues against buying nine, and against assuming a hub-free set of sensors is a smaller version of a hub-based one. It is a different product doing a narrower job.
What about glass break and presence?
Both are on this page because both are routinely shopped as hub-free alternatives to a contact sensor, and neither is a substitute for one.
The Ring Glass Break Sensor uses acoustic detection, which is the correct mechanism for a large pane and covers a whole room rather than a single opening. That drops the per-window cost sharply on a room with three windows in it. It also depends on Amazon Sidewalk, which is a reasonable trade for many households and a firm dealbreaker for anyone building a local-only setup.
The SwitchBot mmWave radar sensor detects a person who is sitting still, which every passive infrared sensor is bad at. It is an occupancy sensor for automation, and it is honest about that: mmWave sees through thin partitions, so it is excellent at knowing a room is occupied and poor at telling you who is in it. Do not buy it as a perimeter device.
How we picked
Compiled from manufacturer manuals, published Matter and Thread documentation and vendor support pages, with no bench measurement of any kind and no claim to have handled the hardware. Our methodology publishes the rubric. If you are choosing an ecosystem as much as a sensor, the Ring versus eufy comparison covers what each one does with your device data, and a sensor purchase usually turns into a smart lock purchase within a month.
How to choose when you do not want infrastructure
Answer one question first: alert here, or alert there?
If the useful response happens at the door, buy a standalone sensor with a local alert and stop. If the useful response happens on your phone while you are elsewhere, you need a hub, a border router or a shared network, and the only real decision is which. There is no product that does the second thing with nothing.
Audit what is already plugged in
Walk the house and list the mains-powered smart devices. An Apple TV 4K, a HomePod mini, a Nest Hub (2nd gen) or a SmartThings Station changes the answer completely, because it turns the Aqara P2 from a hub purchase into a drop-in device with the best local behavior on the page. This audit takes two minutes and routinely saves the price of a hub.
Read the Sidewalk trade-off deliberately
Amazon Sidewalk is a low-bandwidth shared mesh built from Echo and Ring devices in your area. Using it means no base station, which is genuinely convenient, and it means your device's connectivity depends on a network operated by a third party and enabled or disabled by its participants. That is a legitimate choice. It is not the same choice as a local-only setup, and a page that lists it as simply “no hub required” has skipped the part you needed.
What decides the price
Standalone devices carry the cost of their own alerting hardware, which is why a sensor with an onboard alert costs more than a bare Zigbee endpoint that costs almost nothing but does nothing alone. Radar and acoustic sensing cost more again because they need real signal processing rather than a reed switch. Bluetooth is the cheapest radio, and it is cheap precisely because its range is short.
The mistake buyers make
Buying a Thread sensor because the listing says no hub required, then buying a hub anyway. It is the most common return in this category and it is entirely avoidable with a two-minute check of what is already on your shelf. The second most common is buying a standalone Bluetooth sensor for a detached garage sixty feet from the house and expecting the phone to reach it. For anything past the walls of the main building, range is the spec, and only a long-range radio solves it.
If your real requirement is a whole-house perimeter rather than one door, stop reading this page and use the main door and window sensor roundup instead. A hub amortized across nine openings is cheap; hubless sensors on nine openings are not.