What if your smart home knew someone was in the living room without a single lens pointed at them? That is the promise of camera-free presence detection in Home Assistant, and it is easier to build than most people expect. Instead of filming a room, you can lean on motion sensors, radar, Bluetooth signals, and even your own phone’s connection to the Wi-Fi network. Each method has trade-offs, and the wrong mix will leave lights flicking on and off at random. This guide walks through the practical options, the mistakes that trip up beginners, and how to combine sensors so rooms respond reliably while your privacy stays intact.
Why skip cameras in the first place?
Cameras are the obvious choice for presence detection, but they bring baggage. They generate constant video streams, they need careful placement to avoid recording neighbors or family members in private moments, and in many regions they raise data-protection questions. Under the General Data Protection Regulation, video footage of identifiable people counts as personal data, which means obligations around consent, storage, and retention. For a home automation setup that only needs to know “is someone here,” that is a lot of overhead for a simple yes-or-no answer.
Camera-free detection sidesteps those concerns. It also tends to be lighter on your network and your hardware budget. A handful of motion sensors and a Bluetooth proxy can cover a whole floor for less than the cost of one good indoor camera. The trade-off is that no single sensor tells the full story, so the real skill is combining them well. If you already run cameras elsewhere and want to keep them out of your automation logic, the same principles apply to the best home security cameras 2026 setups where detection and recording are handled separately.
The core building blocks Home Assistant gives you
Home Assistant treats presence as a broad concept. Its documentation describes device trackers as the central mechanism, and those trackers come in several flavors: router-based, GPS-based, and network-based. A router integration can report which devices are connected to your Wi-Fi, which tells you someone is home even if you never touch a motion sensor. The companion mobile app adds GPS and zone tracking, so the system knows when a phone leaves the house and when it comes back.
On top of device trackers, Home Assistant supports binary sensors from motion detectors, door contacts, and occupancy sensors. These feed automations directly. The key insight for beginners is that device trackers answer “is anyone home?” while room sensors answer “is someone in this room?” Mixing up the two is the most common reason automations behave strangely.
Motion sensors: the reliable workhorse
Passive infrared, or PIR, motion sensors are cheap, battery-friendly, and widely supported. They detect changes in heat signatures, which makes them good at catching a person walking into a room. They are also the source of the classic complaint: the lights turn off while you are sitting still reading.
PIR sensors have no idea you are still there once you stop moving. The fix is not a better PIR sensor; it is a smarter automation. Common approaches include restarting the off-timer whenever motion is detected, combining PIR with a door sensor to know a room is occupied, and adding a manual override switch. Many people also set a generous timeout, such as five or ten minutes, rather than the default one or two. In practice, a well-tuned PIR setup handles hallways, closets, and bathrooms beautifully and struggles only in rooms where people sit still for long stretches.
Radar and mmWave sensors: presence that waits for you
Millimeter-wave radar sensors changed the game for rooms where people sit. Unlike PIR, these sensors detect the tiny movements of breathing and small shifts in position, so they keep reporting presence even when someone is motionless on a couch. Popular options in the community include the Aqara FP2 and open hardware like the Everything Presence line, and they integrate with Home Assistant either natively or through local bridges.
The catch is cost and tuning. Radar sensors are pricier than PIR, and they need calibration. A poorly placed mmWave sensor will detect through walls, pick up a ceiling fan, or trigger on a pet. Placement matters enormously: mount them at a sensible height, avoid pointing them at reflective surfaces, and use the sensitivity and zone settings to carve out the area you actually care about. When set up well, radar is the closest thing to “the room knows I am here” without a camera.
| Sensor type | Detects still people? | Typical cost | Best for |
|---|---|---|---|
| PIR motion | No | Low | Hallways, closets, quick pass-throughs |
| mmWave radar | Yes | Medium to high | Living rooms, offices, bathrooms |
| Door contact | Indirectly | Low | Entryways, confirming occupancy |
| Bluetooth/Wi-Fi tracking | Yes, if device present | Low | Whole-home “someone is here” logic |
Bluetooth and Wi-Fi tracking for room-level awareness
If you carry a phone or a smartwatch, you are already broadcasting a signal. Bluetooth Low Energy tracking uses that signal to estimate which room you are in. The common recipe is a set of ESPHome Bluetooth proxies placed around the house, feeding a trilateration layer that estimates distance from each proxy. Projects like Bermuda and ESPresense handle the math, and Home Assistant exposes the result as a sensor you can automate on.
Wi-Fi tracking works on a similar principle but at a coarser resolution. Your router or access points report signal strength, and Home Assistant can estimate proximity to each one. It is less precise than Bluetooth and much noisier, because Wi-Fi signal strength fluctuates with walls, furniture, and even microwave ovens. The community consensus is that RSSI-based room detection needs heavy calibration and should be treated as a hint rather than a fact. It works best for “is this person home at all” logic, not “which chair are they in.”
Give each family member a dedicated BLE tag rather than relying on their phone. Phones aggressively manage Bluetooth to save battery, which causes dropouts. A small tag clipped to a bag or keys is far more consistent.
Combining sensors into a presence engine
The real magic happens when you stop treating any single sensor as the truth. A robust room presence setup usually layers three things: a fast trigger (PIR or door contact), a slow confirmer (mmWave or BLE), and a timeout. The trigger turns the lights on instantly; the confirmer keeps them on; the timeout turns them off after everyone has genuinely left.
A few practical rules that save hours of frustration:
- Never let one sensor turn lights off. Require agreement from at least two sources, or a long timeout, before switching off.
- Use door contacts as gates. A door opening is a strong signal that occupancy is about to change.
- Add a manual override. A simple switch or a dashboard toggle lets people say “keep the lights on” without fighting the automation.
- Log your sensors for a week. Home Assistant’s history view reveals false triggers you would never guess at.
- Test at night. Radar and PIR behave differently in the dark, and that is when bad automations are most annoying.
If you want a broader look at how these sensors fit into a full smart-home stack, the guide to the best smart speakers for home automation covers the voice and hub side of the equation, which pairs naturally with presence-driven routines.
Privacy, accuracy, and the honest trade-offs
Camera-free detection is not automatically private, but it is far easier to keep private. A PIR sensor reports “motion: yes” and nothing more. A radar sensor reports presence and sometimes distance, but no image. Bluetooth tracking does reveal that a specific device is nearby, which is a form of personal data, so treat Bluetooth tag IDs as sensitive and keep them local. Running everything on a local Home Assistant instance, with no cloud dependency, is the strongest privacy posture available.
Accuracy is the honest trade-off. Cameras can tell a person from a pet from a shadow with high confidence. Cheaper sensors cannot. That is why the community keeps returning to sensor fusion: no single device is trusted, but the combination is reliable enough for daily life. Expect to spend a weekend tuning, and expect a few weeks of small adjustments afterward. The payoff is a home that responds to you without ever watching you.
For anyone building out a broader automation plan, it helps to think about which rooms deserve radar and which can survive on PIR. Bedrooms and living rooms, where people sit still, justify the extra cost. Hallways and utility rooms rarely do. And if you are weighing subscription-free security hardware alongside your presence sensors, the roundup of the best outdoor security cameras without subscription is a useful reference for keeping everything local. (NIST Cybersecurity Program)
Conclusion
Camera-free presence detection in Home Assistant is a layered craft rather than a single purchase. PIR sensors give you fast, cheap triggers. mmWave radar keeps the lights on when you are sitting still. Bluetooth and Wi-Fi tracking tell the house who is home and roughly where. Door contacts and manual overrides smooth out the edges. None of these is perfect alone, but together they form a system that respects privacy and still feels responsive. Start with one room, log your sensors, tune the timeouts, and expand only when the first room behaves. That patient approach beats any single “best sensor” recommendation, because the best setup is the one tuned to how your household actually moves.
Frequently asked questions
Can Home Assistant detect presence without any cameras at all?
Yes. Home Assistant supports motion sensors, mmWave radar, door contacts, Bluetooth tracking, Wi-Fi-based device trackers, and phone GPS. None of these require a camera, and many run entirely on local hardware.
Which sensor is best for a room where people sit still?
Millimeter-wave radar is the strongest camera-free option for still occupancy because it detects small movements like breathing. PIR sensors will report “clear” once a person stops moving, which is why radar is worth the extra cost for living rooms and offices.
How accurate is Bluetooth room tracking?
Bluetooth room tracking is good but not perfect. It depends on proxy placement, calibration, and whether the tracked device broadcasts consistently. Dedicated BLE tags are more reliable than phones, which throttle Bluetooth to save battery.
Do I need special hardware to run Bluetooth tracking?
You need Bluetooth receivers spread around the house. Many people use inexpensive ESP32 boards running ESPHome as Bluetooth proxies, which forward signals to Home Assistant over Wi-Fi. Trilateration software then estimates which room a tag is in.
Why do my motion-activated lights keep turning off while I am in the room?
PIR sensors only detect movement, so they report “clear” when you are still. Fix this by extending the timeout, combining the sensor with a door contact or radar sensor, or adding a manual override switch that keeps the lights on.
Is camera-free presence detection more private?
Generally, yes. Sensors like PIR and radar report presence without capturing images, and a local Home Assistant setup keeps that data on your own network. Bluetooth tracking still involves device identifiers, so keep those local and avoid cloud integrations where possible.
One last habit worth adopting: name your sensors and automations clearly, and document the logic in a comment or a dashboard card. Six months from now, “binary_sensor.office_2” will mean nothing, but “Office Radar — keeps desk light on” tells you exactly what to check when something misbehaves. Pair that with a simple maintenance routine — swap PIR batteries, re-check radar zones after moving furniture, and revisit timeouts each season — and the setup stays reliable instead of slowly drifting into false triggers. Presence detection is never finished, but a documented, layered system is easy to keep honest.
