Aqara FP2 Presence Sensor: Local Setup in Home Assistant

The Aqara FP2 is a Wi-Fi presence sensor that uses millimeter-wave radar — not a passive infrared sensor — to detect people in a room. It can distinguish up to five people simultaneously and divide a room into up to 30 configurable zones. For anyone running Home Assistant with a local-first network, the obvious question is: does it work without Aqara’s cloud in the loop?

The short answer is yes, with one notable caveat. This guide walks through the full setup: understanding why the FP2’s integration path is different from every Zigbee device Aqara makes, configuring zones before you cut cloud access, pairing the sensor to HA via HomeKit Controller, and locking down outbound traffic without breaking local operation.

What makes the FP2 different from Aqara’s Zigbee sensors

Most Aqara sensors — door contacts, temperature sensors, motion sensors — use Zigbee. You pair them to a coordinator (Zigbee2MQTT or ZHA), and the hub either disappears from the picture entirely or becomes optional. If you’ve done that before, the FP2 will throw you off.

The FP2 is a 2.4 GHz Wi-Fi device. It does not use Zigbee. You cannot pair it to a Zigbee coordinator, and it will not show up in Zigbee2MQTT no matter how long you wait. The integration path into Home Assistant is via the HomeKit Controller integration — HA acts as a HomeKit controller directly over your LAN.

This matters for two reasons:

First, you don’t need Aqara’s hub (M2, M3, E1) to use the FP2 in HA. The sensor speaks HomeKit natively. No middleman.

Second, you don’t need Apple hardware. HomeKit Controller in HA does not require an Apple ID, an Apple Home app, or an Apple hub. “HomeKit” here refers to the protocol — a well-documented local LAN protocol that HA implements independently.

If you’re also running Aqara’s Zigbee devices, the FP2 is the Wi-Fi exception to your usual coordinator workflow — it sits outside Zigbee entirely.

The FP2 requires permanent mains power via a USB-C cable. It does not run on batteries. Plan your placement around a power source.

What “local” actually means for the FP2

Before firmware 1.1.6_0005.0025, the FP2 needed an active internet connection to function — presence detection stopped working when cloud access was cut. Firmware 1.1.6 changed that. After the update, the device operates fully locally for HA automations. Day-to-day presence detection, zone state reporting, and entity updates all happen over the LAN without any Aqara server involvement.

Here’s the honest dependency map:

The practical implication: do your zone planning before you block internet access. Once zones are set and saved in the app, they persist on the device. They appear in HA as individual entities after pairing. But if you later want to resize a zone or add a new one, you’ll need to temporarily restore the device’s internet access, open the app, make changes, and cut access again.

This is the one remaining cloud dependency worth planning around. It’s not a daily concern, but it matters if you’re the kind of person who iterates on zone layouts.

Initial setup — configuring zones in the Aqara Home app

Before pairing the FP2 to Home Assistant, you need to configure zones in the Aqara Home app. Android and iOS both work — you don’t need an iPhone.

A few things to think through before you open the app:

Name zones with HA in mind. Whatever you name a zone in the Aqara app becomes the entity name in Home Assistant. “Zone 1” is not useful. “Desk,” “Sofa,” “Bed,” “Door” — names that map to actual locations in the room make your automations readable without constant renaming later. You can rename entities in HA after pairing, but setting sensible names up front is worth the minute it takes.

Start with fewer zones than you think you need. The sensor supports up to 30 zones, but a cluttered zone map is harder to maintain and produces more noise in your automations. Most rooms work well with four to eight zones covering the areas you actually care about. You can always add more later — but adding more means re-opening the app and cloud access.

Run the calibration. The app walks you through a room calibration that maps the sensor’s detection area. Run it with the room in its normal configuration — furniture in place, people absent. Redoing calibration later requires the app.

Once zones are saved and the sensor is reporting correctly in the Aqara Home app, you can close the app. The zones are stored on the device, not in the cloud. The app is no longer required for daily operation after this point.

Pairing the FP2 to Home Assistant via HomeKit Controller

Prerequisites:
– The FP2 must be on the same network segment as your Home Assistant host, or at minimum your HA host must be able to reach it over the LAN (important if you’re using VLANs — see the next section).
– The HomeKit pairing code — an 8-digit number printed on the sticker on the back of the device, in the format XXX-XX-XXX.
– Pairing mode: if the device has already been used, you’ll need to reset it to put it back into pairing mode. Press and hold the reset button on the device until the LED blinks rapidly.

In Home Assistant:

1. Go to Settings > Devices & Services.
2. Click Add Integration.
3. Search for and select Apple or HomeKit Controller (the integration is listed under both).
4. HA will scan for discoverable HomeKit devices on your LAN. The FP2 should appear as an unpaired device.
5. Enter the 8-digit pairing code from the device’s sticker.
6. Complete pairing.

After successful pairing, HA creates a device with multiple entities:

• One binary sensor per zone (presence detected / not detected)
• One overall presence binary sensor for the whole detection area
• One illuminance sensor (lux)

The zone entities are named based on whatever you set in the Aqara app. The overall presence sensor is particularly useful for room-level automations — “any presence in living room” — while individual zone sensors let you do things like distinguish “presence at desk” from “presence on sofa.”

If HA doesn’t discover the device automatically, check that the FP2 and your HA host are reachable to each other and that mDNS or Bonjour is functioning on your network. In some VLAN setups, mDNS traffic needs to be explicitly allowed or relayed.

Blocking cloud traffic without breaking local operation

The FP2 phones home to Aqara’s servers when it has internet access. You don’t need to trust that communication if you’d rather not. After initial setup and pairing, the device functions entirely on the LAN — no cloud traffic is required.

VLAN placement

The standard approach: put the FP2 on an IoT VLAN with LAN access allowed and WAN access blocked. The device needs to reach your HA host (typically on your main or home VLAN) for HomeKit Controller to function. It does not need to reach any external server.

A representative firewall ruleset, in plain terms:
– Allow: IoT VLAN → Home VLAN on port 5353 (mDNS) and the HomeKit pairing port (51827 by default)
– Allow: IoT VLAN → Home VLAN for established/related return traffic
– Block: IoT VLAN → WAN (all traffic)
– Block: IoT VLAN → all other VLANs

If you’re running HA on its own host on the home VLAN, adjust the allow rules to point at your HA host’s IP. If HA is in a VM or container with its own network interface, make sure the firewall rules account for that interface.

What to expect after blocking WAN

HA’s HomeKit Controller connection to the FP2 continues to work. Entity state updates arrive normally. Automations trigger as expected.

The Aqara Home app will show the device as offline when you open it — that’s expected. You’re not using the app for daily operation. The device is not offline; it’s just unreachable from Aqara’s servers, which is exactly what you want.

If you need to reconfigure zones at some point, temporarily restore the device’s internet access, open the app, make your changes, then block WAN access again.

Verification

One way to confirm local operation: block the device’s WAN access, restart Home Assistant, and watch whether the FP2 entity comes back online and reports presence correctly. If it does — and it should — the device is communicating purely over your LAN. A network capture on the IoT VLAN interface showing no outbound traffic to external IPs is more definitive confirmation, if you want it.

Caveats and known issues

A few things worth knowing before you commit to a zone layout:

Zone editing is cloud-gated. This is worth repeating because it catches people off guard. If you block internet access and then decide your zone layout is wrong, you’ll need to restore WAN access temporarily to fix it. Plan carefully during initial setup.

Entity naming is set at pairing time. The HA entity names for your zones are determined by what you named them in the Aqara app before pairing. After pairing, you can rename entities in HA’s entity registry without affecting function, but the underlying device name and entity ID are derived from the app names. If the app names were “Zone 1,” “Zone 2,” you’ll be living with those slugs in your entity registry.

mDNS matters on segmented networks. If your HA host and FP2 are on different VLANs, mDNS discovery won’t work by default. You’ll need an mDNS reflector or repeater (Avahi on the HA host can handle this in many setups). Without it, HA can’t discover the device during pairing — you’d have to specify the IP directly, which HomeKit Controller doesn’t natively support in a straightforward way.

The FP300 is the non-Wi-Fi alternative. If you want a presence sensor but prefer to keep Wi-Fi devices off your IoT VLAN, Aqara released the FP300 — a combined PIR and mmWave presence sensor. It ships with Thread/Matter firmware by default and also supports Zigbee via an alternate firmware image; in Zigbee mode it integrates through Zigbee2MQTT like most other Aqara Zigbee devices, with no Aqara hub needed if you already run a coordinator. If you’ve already built a Zigbee network, the FP300 is worth considering. The FP2 remains the better option if you want mmWave zone mapping without building out a Zigbee coordinator, or if you need the FP2’s larger zone count.

Is the FP2 the right local presence sensor for your setup?

The FP2 makes a strong case for rooms where you want reliable presence detection — spaces where a PIR sensor generates false negatives because people are sitting still. Living rooms, offices, bedrooms all benefit from mmWave. PIR sensors are still fine for hallways, doorways, and areas where motion is the relevant signal.

For readers choosing between sensor types, the tradeoff is straightforward: PIR sensors react to motion and miss a still person, while mmWave radar holds presence on someone sitting motionless. Use the right one for the room.

The cloud-only-for-setup model is reasonable. You spend one session online — update firmware, configure zones, pair to HA — and then the device operates independently. The zone-editing caveat is real but manageable if you plan ahead. Most home setups don’t require frequent zone reconfiguration.

The HomeKit Controller integration is reliable and well-maintained in HA. You don’t need any custom component, HACS repository, or third-party bridge — it’s a first-party HA integration that works with standard firmware.

If you’re already running an Aqara hub for other devices, the hub path is an alternative integration method — see our Aqara Hub M2 local setup guide for how that works — but for the FP2 specifically, the direct HomeKit Controller path is simpler and eliminates the hub as a dependency.