Skip to content

New: HelloTime now integrates with HelloBooks — sync clients, hours, invoices and payroll.

Mobile attendance

Early access · pilot customers only

Face-recognition attendance, anti-spoof, sub-second match.

Workers clock in by looking at the front camera. HelloTime runs MediaPipe liveness on the device, verifies the match against Azure Face API, and writes the clock event in under 1.2 seconds — with a manager-approval queue, PIN fallback, and a thumbnail-only retention policy that keeps India DPDP and EU GDPR teams happy.

The end-to-end Azure Face enrol and verify loop is rolling out to pilot customers while we finish wiring the backend endpoints — request a pilot slot.

How face-recognition attendance works in HelloTime

Currently shipping to pilot customers. The same template will power the phone clock-in screen and the tablet kiosk on every site, so you enrol a worker once and they show up everywhere.

  1. 01

    Enrol once, in three angles

    At hire, the worker captures three reference angles from the front camera. Frames are sent to Azure Face API and stored as a PersonGroup template plus an encrypted thumbnail in Azure Blob — the raw photo is never retained on the device.

  2. 02

    Look at the camera to clock in

    The HelloTime mobile app opens the front camera in a circular viewfinder, with on-device MediaPipe liveness running before any frame leaves the phone. A single live frame is matched against the worker's template via Azure Face Verify.

  3. 03

    Sub-second match, sub-1.2-second clock event

    End-to-end latency target is under 1.2 seconds from tap to confirmed clock event, including geofence validation and idempotent server write. The kiosk path uses the same pipeline with a per-site PersonGroup cache for sub-second identification.

  4. 04

    Anti-spoof escalation

    If liveness fails or the match score falls below threshold, HelloTime forces a fresh selfie plus a manager-approval queue entry. Repeated failures escalate to PIN + selfie fallback so a clock event is never silently dropped on the worker.

Anti-spoof, liveness and the failure path nobody else ships.

Buddy-punching is the #1 deskless attendance fraud. HelloTime treats every borderline match as something a human, not the app, should resolve.

On-device passive liveness

MediaPipe blink and head-turn checks run on the phone before any pixel is uploaded — printed photos and static screens never get past the device.

Active challenge on flagged events

If a frame looks borderline, the app asks for a blink or a left-right head turn. This kills the printed-photo and looped-video replay attacks Truein customers complain about most.

Manager-review queue

Below-threshold matches are not silently approved. They land in a manager-review queue with the timestamped frame, so a human, not the app, breaks the tie.

Fallback path

PIN plus selfie keeps the day moving when lighting or hardware fail — no clock event is lost because a camera couldn't focus.

Why HelloTime beats Truein on face attendance.

Truein owns the face-attendance category in India today. Here is how HelloTime is built to take it.

Sub-1.2s end-to-end latency

HelloTime pairs Azure Face Verify with on-device liveness to keep the full clock event under ~1.2s on a ₹15k Android tablet — Truein reviews on G2 routinely flag multi-second delays at shift start.

Same template, phone and kiosk

One PersonGroup serves the worker's phone clock-in and the site's tablet kiosk — no duplicate enrolment, no separate face DB. Truein's kiosk and mobile flows do not share a single recognition path.

Privacy by default

Raw photos are never stored on the device. Templates live in Azure Face; thumbnails live in Azure Blob with Hot → Cool → Archive lifecycle tiering. India DPDP-aware from day one.

Public, self-serve pricing

Face attendance is part of HelloTime's published Attend SKU. No sales call. Truein gates pricing behind a demo.

Privacy and storage

  • Liveness checks run on the device. Frames that fail liveness are never uploaded.
  • The verified frame is sent to Azure Face API; HelloTime stores a PersonGroup ID and an encrypted thumbnail in Azure Blob — never the raw photo.
  • Thumbnails follow a Hot → Cool → Archive lifecycle (30 days, 1 year, then archive) to keep storage cost negligible per seat.
  • India DPDP Act and EU GDPR aware: workers can request export or erasure; deletion removes the PersonGroup entry and the thumbnail in one step.
  • Manager-approval queue, fallback PIN+selfie, and admin audit log are all on the published Attend SKU — no add-on.

Start tracking time in under 2 minutes.

Free forever for up to 5 employees. No credit card. Paid from ₹49/user/month.