Local-first AI health platform

Turn a normal webcam into a health and wellbeing platform.

HealthCam runs camera-based analyzers locally, connects them to cloud dashboards, and turns webcam signals into modular health and wellbeing features.

Started as a validated B.Sc. thesis prototype for camera-based BPM estimation, HealthCam is evolving into a local/cloud platform for computer vision health analyzers, plugin modules, telemetry pipelines, and dashboard-ready insights.

Local-first AIComputer visionHealth telemetryPlugin modulesCloud dashboards

Today’s wellbeing session

72 min focused

local camera

Screen presence

Processed

35 min in front of screen

Suggested action

Break recommended

Take a 2-minute eye break

Example analyzer

rPPG / BPM

Processing mode

Local-first

Module store concept

Add new health and wellbeing analyzers as the platform evolves.

Platform overview

Local analysis, cloud product experience, installable modules.

HealthCam is designed as a platform rather than a single analyzer. The local machine handles sensitive and hardware-close processing, while the cloud provides dashboards, configuration, and long-term product experience.

Local machine

Captures webcam input, runs local services, and executes computer vision analyzers close to the user’s hardware.

  • Webcam capture
  • Local analyzers
  • Service orchestration

Cloud dashboard

Handles authentication, telemetry storage, insight tables, dashboards, configuration, and product experience.

  • Authentication
  • Telemetry storage
  • Dashboard insights

Plugin ecosystem

Allows new analyzers to become installable modules with their own service, telemetry schema, transformations, and dashboards.

  • Plugin manifest
  • SQL transformations
  • Module dashboards

Modules

From research validation to practical wellbeing tools.

HealthCam started by validating camera-based BPM estimation. The next step is turning that foundation into useful modules for everyday computer wellbeing.

Validated foundation

Camera-based BPM research

HealthCam started from a working rPPG thesis prototype that estimated heart rate from normal webcam input under real computer conditions.

Building now

Eye-strain support

A practical wellbeing module designed to suggest breaks based on actual screen presence rather than fixed timers alone.

Building now

Presence-aware sessions

HealthCam can distinguish between open-laptop time and actual time spent in front of the screen.

Platform direction

Stress-pattern awareness

Future modules may combine physiological signals, presence, and work-session patterns to help users notice stress and recovery trends.

How it works

A complete path from webcam signal to dashboard insight.

HealthCam connects local computer vision services to cloud-side telemetry processing, allowing analyzer outputs to become product features instead of isolated research scripts.

1

Camera input

A normal webcam captures signals from the user’s environment.

2

Local processing

Computer vision analyzers run locally whenever possible, close to the hardware.

3

Telemetry pipeline

Analyzer outputs are sent as structured telemetry into the cloud pipeline.

4

Dashboard insights

SQL transformations generate dashboard-ready insights from raw telemetry.

Processing mode

Local-first

Camera input

Handled on the user device where possible.

Analyzer execution

Runs as local services close to the hardware.

Cloud dashboard

Used for visualization, settings, and product experience.

Privacy direction

Camera data should stay close to the user.

HealthCam is designed around local-first processing because camera input can be sensitive, latency-dependent, and hardware-specific. The cloud supports the product experience, but raw webcam processing does not need to be centralized by default.

Platform status

From validated prototype to first user-facing modules.

HealthCam already supports the core local/cloud flow: webcam capture, independent local services, analyzer execution, telemetry ingestion, data processing, and dashboards. HealthCam 3.0 upgrades that foundation into a plugin-based platform so new health and wellbeing analyzers can become installable modules.

Already working

  • Local webcam capture
  • rPPG/BPM analyzer integration
  • Independent local microservices
  • Automatic service lifecycle management
  • Cloud telemetry ingestion
  • Raw telemetry storage
  • Dashboard data processing
  • End-to-end dashboard flow

3.0 platform upgrade

  • Plugin-based analyzer system
  • Manifest-based service discovery
  • Plugin-compatible orchestration
  • Module-specific telemetry schemas
  • Installable analyzer modules

Next product modules

  • Eye-strain break module
  • Presence-aware work sessions
  • First user testing
  • Additional wellbeing analyzers
  • Plugin marketplace direction

Engineering

Built as a real local/cloud computer vision platform.

Behind the product interface, HealthCam uses a local agent, independent analyzer services, service lifecycle management, manifest-based discovery, raw telemetry storage, SQL transformations, and cloud dashboard integration.

Founder note

Built by Diego Turri as a founder-engineer project.

I started HealthCam as my B.Sc. thesis prototype and continued evolving it into a modular platform for camera-based health and wellbeing analyzers. The goal is to make research-level computer vision easier to test, integrate, and turn into useful products.