Cloud Commons Labs · Project 01

LEON

Local Environment Observability Node

A mobile observability engine that detects and classifies behavioural data flows from mobile applications, giving Canadians a device-level view of how their apps participate in the personal data economy.

Built by Cloud Commons Labs. All analysis happens on-device via local VPN interception: no payload inspection, no data transmitted without consent. Opt-in structural profiles contribute to a national picture of Canada's app-layer ecosystem.

Invest in LEON → Grant inquiry →

Stage

Phase 0

Platform

Android First

Data Model

On-Device · No Payload

Pipeline

CCC Contributor

THE FIRST
CIVIC
OBSERVATORY.

The personal data economy has an observation problem. Platforms that operate advertising and analytics infrastructure can see exactly what behavioural data flows where: who generated it, where it went, and what it was worth. Users, researchers, and regulators cannot. LEON explores whether a mobile device can serve as an independent observatory on the same flows.

LEON uses a local VPN interception architecture on Android to capture outbound network traffic metadata, map it to known data infrastructure, including advertising networks, analytics platforms, and third-party SDKs, and classify behavioural data flows in real time. It correlates these network observations with device sensor access to produce a behavioural data profile of each application. No payload inspection. No data leaving the device without consent.

The result is not a score or a warning. It's an instrument. One that surfaces the same structural picture of the personal data economy that platforms have always had, returned to the person the data belongs to.

REAL-TIME
ALERTS FROM
YOUR DEVICE.

LeonBark is the user-facing output of the observability engine. When the local VPN proxy detects a flow worth surfacing, such as a tracker contact, a sensor access event, or an anomalous background call, it fires an immediate on-device alert with enough context to understand what happened. Examples:

🐾

TikTok contacted tracker.com 6× in 60 seconds while running in the background. Tracker classified: advertising/analytics. Jurisdiction: US.

Instagram accessed your microphone at 2:03 AM while your screen was off. No foreground activity detected. Added to weekly digest.

🐾

3 apps made outbound calls during device sleep in the last 24 hours. View per-app breakdown →

📍

Weather app requested precise location 14 times today. Only 2 requests occurred while app was open. Consider restricting to "While Using."

All classification is on-device using public tracker lists (EasyList, EasyPrivacy). No traffic content is read or stored. LEON sees who an app is calling, not what it says.

PHASED.
DELIBERATE.
DEPLOYABLE.

LEON is built in clearly sequenced phases, each with defined acceptance criteria. Phase 0.5 answers the core research question: can a phone reliably observe and classify its own behavioural data flows without inspecting encrypted payloads? Each phase unlocks the next.

// Phase 0 ·2–3 Weeks Dev

Privacy Checkup MVP

Goal: Show the feeling of sovereignty in <10 minutes

On-device audit of app permissions and ad tracking settings. Plain-language exposure report. Duolingo-style streaks. No account required. Ships fully offline.

  • 8–10 one-click privacy switches with OS deep links
  • Data Value estimator (simplified)
  • Android (Kotlin) · SQLCipher local storage
  • Acceptance: Checkup complete in <7 min · works offline

// Phase 0.5 ·3–5 Weeks Dev

LeonBark: Local VPN Proxy

Goal: Show real traffic insights without exfiltrating data

Android local VPN proxy captures per-app domain calls. On-device classification using public tracker lists. Real-time LeonBark alerts. Per-app dashboard.

  • VpnService + OkHttp/pcap parser · no root required
  • Ad/analytics/CDN classification on-device
  • All analysis on-device · no cloud required
  • Acceptance: <5% battery overhead · ≥80% tracker match rate

// Phase 1 ·4–6 Weeks Dev

Wizard Automations + Data Value

Goal: Turn privacy best-practices into one-click routines

Privacy routines ("Secure Me Now"). Broker opt-out launcher for Canadian + US data brokers. Camera/mic activity monitor. Data Value v1.

  • LeonBark: Camera/Mic Activity Monitor
  • Cross-check app foreground state for background access alerts
  • Data Value: tracker calls × CPM proxy = monthly value retained
  • Acceptance: 70% of routines complete in ≤60 seconds

// Phase 2 · Pipeline Contribution

National Map

Goal: Aggregate into public-interest Canadian data commons

Opt-in anonymous structural profile contribution to Cloud Commons Canada. Aggregated into national exposure map published openly. No paywall.

  • No personal data retained on contribution
  • Shape of ecosystem from one more Canadian device
  • Certified by CCC before pipeline write access granted
  • Output: open national app-layer infrastructure map

ONE SIDE
HAS AN
OBSERVATORY.

38M+

Canadians with no independent view of their app data flows

80+

Average apps per Canadian smartphone, most transmitting telemetry

0

Civic observatories of Canada's personal data economy

Platforms that operate advertising and analytics infrastructure maintain the only comprehensive observatories of behavioural data flow. They can see what data is collected, by whom, at what scale, and to what end. Users, researchers, and regulators operate blind. LEON is a prototype attempt to build an observatory on the other side.

Federal privacy law reform (Bill C-27), the CPPA, and sustained media attention to data practices have created real regulatory and public appetite for transparency infrastructure. LEON addresses a gap that no existing Canadian tool fills: built in Canada, analysed on-device, governed in the public interest.

INVESTMENT
+ GRANT
PATHWAYS.

LEON has two distinct funding pathways that are not mutually exclusive.

Equity Investment

Investor Path

Commercial privacy tool with freemium model, API licensing for researchers and institutions, and a data-value product layer. Labs structure supports equity investment and co-founder arrangements.

Grants

Grant Path

LEON's public-interest data contribution function creates genuine eligibility for NRC IRAP, CDMN, digital-sovereignty grants, and civic-tech funding streams. CCC's non-profit status provides additional credibility for public-interest funders.

Labs' partnership with Cloud Commons Canada and the operating principles in the inter-entity agreement are designed to make mission drift structurally difficult, not just culturally discouraged. For funders who care about where their money ends up, that structure is worth understanding.

Read the governance structure →
Get in touch about LEON →