Built for schools that cannot send student data to external AI.
EdSecure is designed around local processing, school-approved evidence and clear technical boundaries.
Trust model
We bring the AI to your data.
The local-first model supports analysis on an approved laptop, local appliance or school-controlled environment. Local AI agents, deterministic rules and document parsers analyse evidence without needing external AI APIs.
- 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
- Written scope before technical testing
Designed boundaries
EdSecure does not require full administrative access to start. The assessment can begin with public signals and school-approved exports, then increase confidence with read-only connectors or scoped validation.
LocalModels, rules and evidence parsing can run inside a controlled environment.
Evidence-ledFindings are separated into confirmed, likely, possible or needs evidence.
Progressive visibility
Schools choose the evidence level.
| Mode | Data handling | Best for |
|---|---|---|
| External Snapshot | No sensitive uploads required | Early discovery and low-friction entry |
| Evidence Upload | Files processed locally | Core assessment and board reporting |
| Read-only Connector | School-approved read-only collection | Repeatable high-confidence assessment |
| Scoped Validation | Written approval for defined tests | Confirming specific leak paths |