Milesight VS121 vs VS133: AI vision or ToF for people counting?

Short answer: Specify the Milesight VS133-P when the space must contain no lens at all — it senses depth with an invisible 940 nm time-of-flight beam, collects no images and works in complete darkness, quoting counting accuracy up to 99.8% over up to four counting lines. Specify the VS121 when you need to know how a whole area is used rather than how many people crossed a threshold: it is an AI sensor specified to count and measure area utilisation without collecting images, mounts at 2.3 m to 4 m and covers up to 6.8 × 18.8 m. Both list GDPR under approvals. The VS133-P costs roughly 1.8 times as much as the VS121.

Which one do I actually need, the VS121 or the VS133-P?

Choose the VS121 if the question you are trying to answer is “how is this space being used?” — an open-plan floor, a meeting suite, a study area. It covers 2.6 × 8.6 m at 2.3 m mounting height, 4.8 × 14.0 m at 3.0 m and 6.8 × 18.8 m at 4.0 m, with a 190° × 112° field of view. You can map up to 16 detection regions of up to 10 sides each with dwell time, plus people-flow analysis at up to four edges. One unit reports a zone, not a single seat.

Choose the VS133-P if the question is “how many people came through here?” — an entrance, a corridor, a gate line. Its detection range is 0.5 to 3.5 m, which caps installation height at 3.5 m, and its 98° × 80° field of view is deliberately tighter than the VS121’s: you want the doorway, not the room behind it. It gives up to four bi-directional counting lines, up to four detection areas, motion and dwell heat maps, ±3.5 cm distance measuring accuracy, and stores up to one million records locally with CSV export.

Is the VS121 a camera in the ceiling?

This is the objection that stalls occupancy projects, so answer it precisely. The VS133-P is not a camera. Its sensing element is a 940 nm infrared beam that is invisible to the eye and measures distance, which is why the datasheet can claim operation in complete darkness. There is no imaging path to discuss in a staff consultation.

The VS121 is not a CCTV camera either — it is specified to count people and measure area utilisation without collecting images, and it supports up to eight privacy masks. Both units carry GDPR alongside CE, FCC and RoHS. The honest distinction is that the VS121 is an optical AI sensor and the VS133-P is a depth sensor, so if your policy or your works council rules out lenses as a category, only the VS133-P removes the argument entirely. Note also that the specifications state darkness operation for the VS133-P only; they give no minimum light level for the VS121 either way.

What the datasheets catch people out on

• Ceiling height decides it before privacy does. The VS133-P caps at 3.5 m. Above that it is out, and the VS121 — rated to 4 m — is the only one of the two that fits.
• Neither is battery powered. The VS121 takes 5 V / 3 A over USB Type-C at 9.4 W, so it needs a supply at the mounting point. The VS133-P takes 802.3at PoE+ at 9.4 W typical and 23.8 W maximum, so it needs a PoE+ switch port instead.
• Both are indoor-only. IP30 for the VS121, IP40 for the VS133-P. Operating temperature is −20 °C to +60 °C and −20 °C to +50 °C respectively.
• They connect differently. The VS121 runs LoRaWAN Class C, OTAA or ABP, at −137 dBm sensitivity, on EU868 for the UK, and also lists Milesight D2D. The VS133-P has no radio: it reports over its Ethernet port using HTTP(S), MQTT(S) or BACnet/IP.

Milesight VS121 AI Workplace Occupancy Sensor

VS121

£276.82 – £284.30Price range: £276.82 through £284.30 ex VAT (£332.18 – £341.16 inc. VAT)

Installation height and detection range2.3 m to 4 m; 2.6 × 8.6 m at 2.3 m up to 6.8 × 18.8 m at 4 m

View the VS121

Milesight VS133-P AI ToF People Counting Sensor

VS133

£492.30 ex VAT (£590.76 inc. VAT)

Detection range (and maximum installation height)0.5 to 3.5 m; installation height ≤3.5 m

View the VS133