Australian • Local-first • School-safe

Find where student data could leak.

EdSecure.ai uses local AI agents to assess AI tools, EdTech vendors, cloud sharing, accounts, devices, networks and integrations — without sending school data to external AI.

No external AI APIs requiredEvidence-backed findingsBoard-ready reportsBuilt for Australian schools
Local Exposure Agent

Student Data Exposure View

Action required

12 confirmed risks • 18 review items • 6 quick wins

72
5High priority leak paths
14AI / EdTech review items
9Cloud sharing fixes
3Board decisions
School systemsLocal AI agentsRisk report

Runs locally

Analysis can run on an approved laptop, local appliance or school-controlled environment.

No student data to external AI

Designed so school evidence is not sent to ChatGPT, Claude, Gemini or other external AI APIs.

Evidence-based findings

Findings are linked to evidence, confidence, student-data impact and recommended action.

The problem

Student data now moves through a messy school ecosystem.

AI tools, EdTech apps, cloud drives, staff accounts, devices, networks, browser extensions and integrations all create exposure points. Most schools do not have one clear view.

Unapproved AI tools

Staff and students may use public AI tools with student work, reports or identifying details.

Vendor uncertainty

Privacy policies, AI-training terms and subprocessors are hard to interpret.

Over-shared files

Student reports or learning support notes may be shared beyond intended groups.

Weak account controls

Missing MFA, dormant accounts and excessive admin access can create leak paths.

Hidden integrations

SSO, OAuth, LMS and vendor integrations may have broader access than expected.

Board visibility gap

Leadership often receives technical noise instead of decision-ready risk priorities.

How it works

Local, scoped and evidence-led.

Scope

Agree systems, evidence sources and boundaries.

Collect

Use public signals, approved exports, policies and vendor documents.

Analyse

Local AI agents and rules analyse evidence on controlled hardware.

Prioritise

Rank by severity, confidence, student-data impact and effort.

Report

Deliver board, principal and IT-ready outputs.

Progressive visibility

Start with limited access. Increase confidence as evidence increases.

ModeAccess requiredWhat it providesConfidence
External SnapshotPublic information and authorised observationWebsite, DNS, email, public exposure and policy/vendor signalsLow–medium
Evidence UploadSchool-approved exports and documentsAI, vendor, cloud sharing, identity, account and policy findingsMedium–high
Read-only ConnectorApproved read-only accessAutomated high-confidence reassessment and trend reportingHigh
Scoped ValidationWritten approval for specific testsEvidence that selected leak paths are real or closedVery high
Example findings

Plain-English risks a school can act on.

HighConfirmed

Privileged accounts without MFA

Accounts with elevated access may expose systems containing student information.

HighConfirmed

Student files shared externally

Assessment or learning-support folders may allow access outside the school domain.

MediumReview

AI tool terms unclear

Vendor terms do not clearly rule out training or model improvement using uploaded content.

Local-first trust

We bring the AI to your data.

EdSecure is built for schools that cannot send student data to external AI. Analysis can run locally using local models, rules and parsers, with no external AI APIs required.

  • No telemetry by default
  • Evidence can be redacted before analysis
  • Findings cite source evidence
  • Assessment data can be deleted or returned
See local-first model

$ edsecure local-assess --school-approved

✓ Loaded policy pack

✓ Parsed vendor documents

✓ Analysed cloud sharing exports

! 5 high-priority exposure paths found

✓ Board report generated locally

Know where student data could leak before you have to explain it.

Book a discovery call to see whether EdSecure is a fit for your school.