Passive Overdose Detection at the Edge

Miles CAM Overdose Detection helps staff respond faster when someone may have collapsed from an overdose. No wearable, no app, no check in. Built for sensitive environments with privacy first design and a human review step.

Edge
No Cloud Video
< 2s
Signal to Alert
Review
Human Confirmed

The Problem

In shelters, supervised consumption sites, and public washrooms, overdoses can happen unseen. Wearables don’t scale, manual check-ins are inconsistent, and most camera systems aren’t built for privacy-sensitive environments.

Our Solution

Miles CAM Overdose Detection is passive monitoring designed for real sites. It looks for collapse risk signals, routes them to a staff review queue, and escalates only when confirmed so the system stays trustworthy.

How It Works

Three sensing layers, one shared vocabulary, and a safety-first escalation flow

Layer 1: Camera (MVP)

An IP camera streams to a local device. The system finds people in the scene and tracks posture and stillness over time to detect collapse risk.

  • Person detection + pose landmarks
  • Sliding-window posture & stillness checks
  • Works without cloud video

Layer 2: RF breathing signal (Phase 2)

Optional RF sensing can detect tiny breathing rhythm signals that video cannot reliably confirm. This adds a strong second opinion for safety.

  • Breathing rhythm detection
  • Detects slowed or missing rhythm
  • Runs locally

Layer 3: Clip understanding (Phase 3)

Optional clip-level understanding adds scene context over a few seconds of video and improves as your team reviews and labels real events.

  • Short clip context
  • Reduces false alarms over time
  • Improves from review feedback

One decision, multiple signals

Signals from each layer combine into one final risk level. This reduces false alarms and avoids missing critical events.

  • Clear risk levels
  • Simple labels staff can use
  • Confidence combines across signals

Human-in-the-Loop Review

Alerts first enter a staff review queue. Confirm, mark false alarm, or escalate so the system earns trust and stays usable.

  • Blurred / privacy-safe review clips
  • Audit trail of decisions
  • Escalation only when confirmed

Edge + Privacy First

Designed for sensitive environments. Raw video stays on-device; storage is minimal, and privacy protections are built in.

  • No cloud video by default
  • Configurable retention
  • Deployment-ready for harm reduction

Privacy by Design

Built to be deployable in real harm reduction environments

Edge first

Designed so raw video stays on-device by default.

  • No cloud video required
  • Local processing on edge hardware
  • Works offline in constrained sites
  • Configurable retention

Privacy Protections

Reduce privacy risk while maintaining operational usefulness.

  • Face blurring before storage
  • Minimal stored metadata
  • Role-based access (planned)
  • Audit trail for reviews

Human Review

Staff remain in the decision loop to prevent trust loss from false alarms.

  • Review queue before escalation
  • Confirm / dismiss / escalate
  • Clip-level evidence
  • Feedback improves models

Deployable Sites

Where passive monitoring can save a life.

  • Supervised consumption sites
  • Shelters & transitional housing
  • Public washrooms
  • Common areas / intake zones
Phase 1

Roadmap

Ship a camera-only MVP first, then add RF sensing and clip-level understanding

Phase 1: Camera MVP

YOLO + pose estimation + rule engine for collapse risk signals, plus a human review queue and basic alert escalation.

Edge Deployment Focus

Privacy-first architecture designed for on-site processing and sensitive environments.

Pilot Partnerships

Recruiting pilot sites in harm reduction, shelters, and transitional housing to validate the MVP and collect real-world feedback.

Phase 2: RF Breathing Signal

Add ESP32-based RF sensing to detect breathing rhythm changes and strengthen confidence when the camera alone is ambiguous.

Phase 3: Clip level understanding

Add a video transformer layer and continuously improve using labels collected naturally in the review workflow.

Join Beta Program

Be Among the First

Join the pilot and help shape overdose detection that works in the real world

Beta Program Benefits

  • Free access during beta period
  • White-glove setup and installation
  • Direct access to development team
  • Influence product roadmap
  • Early access to new features
  • Priority support
  • Case study opportunities

Ideal Beta Customers

  • Supervised consumption sites
  • Shelters & transitional housing
  • Public washrooms & supported housing facilities
  • Harm reduction and public health organizations

Apply for Beta Access

We'll review your application and get back to you within 48 hours.

Investment Opportunity

Invest in privacy first overdose detection

Miles CAM is seeking strategic investors to accelerate pilots, deployments, and life saving outcomes

Market Opportunity

  • Large unmet need in harm reduction + shelters
  • Passive monitoring beats wearables for adoption
  • Edge privacy unlocks sensitive deployments
  • Scales from overdose into broader safety

What We've Built

  • Camera MVP with posture + stillness signals
  • Human review queue to control false alarms
  • RF breathing layer (Phase 2)
  • Video transformer layer (Phase 3)
  • Edge first privacy architecture

Investment Use

  • Pilot deployments & evaluation
  • Engineering: camera MVP → RF layer
  • Data labeling workflow + model improvement
  • Security, privacy, and production hardening

Target Metrics

  • 6 months: 3 to 5 pilots live
  • 12 months: multi site expansion
  • 18 months: repeatable deployment playbook
  • Year 2: scaled rollouts with partners
  • Year 3: broader safety monitoring expansion

Investor conversations

We are speaking with mission-aligned investors and partners. Contact us for the investor deck and details.

Get in Touch

Pilot sites, partners, and investors. Let’s talk.

Pilot Deployments

For supervised consumption sites, shelters, and facilities looking to pilot overdose detection, contact us directly.

Partnerships

We’re looking for harm reduction partners, operators, and integrators to deploy responsibly and iterate fast.

Send us a Message