Features

AI prompt data protection logo – Trust-Prompt
Trust-Prompt Features v1.12 Local • Deterministic • Enforced

AI prompt data protection before you send

AI prompt data protection is about preventing accidental leakage of sensitive data to AI tools. Trust-Prompt checks what you type locally and enforces a decision (WARN or BLOCK) before transmission.

Pricing Testing & Quality How it works
Install Extension View Pricing Basic vs Pro
Local processing Deterministic rules No server calls No AI models WARN = explicit consent BLOCK = send prevented

StoreBasic: deterministic local checks + enforced WARN/BLOCK for AI prompt data protection on ChatGPT.
Pro: extends AI prompt data protection with multi-site scope controls, upload policy, and settings lock (when licensed).

AI prompt data protection engine (StoreBasic)

Trust-Prompt performs a local and deterministic pre-check on what you type before it can be submitted to supported AI chat sites. It runs offline: no servers, no telemetry, no AI calls.

  • Deterministic rule engine (offline)
  • No server calls, no backend, no tracking
  • No AI models / no AI APIs
  • Enforced before transmission (pre-send)
How it works Testing & Quality
Chat input · Trust-Prompt pre-check
Checking…
Your message (example)
“Can you help me summarize this contract clause?”
Status appears while typing and before submission.

Built for “no surprises”

Because AI prompt data protection must be reliable, Trust-Prompt holds the send action until the decision is resolved. Fast submits cannot bypass the result.

  • Sticky decision state (persists until resolved)
  • Explicit consent path for WARN (Proceed anyway)
  • Submission cannot happen before the decision is handled

AI prompt data protection with BLOCK (high-risk)

High-risk patterns are blocked to prevent accidental leaks.

  • IBAN (validated formats)
  • Credit card combinations (card + CVV / expiry)
  • Private keys and credential material
  • Tokens / API keys / bearer tokens
  • Regional identifiers (where applicable)

BLOCK means: send is reliably prevented until the high-risk content is removed.

BLOCK overlay (screenshot)

Real UI capture.

AI prompt data protection – Trust-Prompt BLOCK overlay screenshot preventing high-risk submission

WARN overlay (UI mockup)

Explicit consent: proceed only if you choose.

Chat input · Trust-Prompt pre-check
WARN
Detected indicators (example)
Email address pattern · Phone pattern · Address-like pattern

Warning — review before sending

This message contains medium-risk indicators. Proceed only via an explicit confirmation.

Proceed anyway Edit message
WARN is never silent. You must choose.

AI prompt data protection with WARN (review)

Warnings surface common personal data indicators and structured signals.

  • Email patterns
  • Phone patterns
  • Address-like patterns (heuristics)
  • Structured exports (CSV/tabular)
  • Document classification markers (where applicable)

Audit for AI prompt data protection (Pro only)

Pro only: Trust-Prompt can provide local audit visibility without storing your prompt content. It shows counts and rule IDs (e.g., how often WARN/BLOCK occurs) for validation and governance.

  • Local counters (WARN / BLOCK / events)
  • Rule IDs only (no prompt text stored)
  • Helps validate deterministic behavior

Availability: Pro (licensed). StoreBasic does not include audit.

Audit panel (UI mockup)

Counts and recent rule IDs (illustrative).

Trust-Prompt · Audit (local)
Pro only No prompt storage
BLOCK events
High-risk submissions prevented
12
Recent rule IDs: FIN_IBAN-001 SEC_TOKEN-003
WARN events
Explicit consent required
37
Recent rule IDs: PII_EMAIL-001 PII_PHONE-002
Upload events
Attachment events detected
5
Upload event detection is available; upload policy controls are Pro-only.
Last check
Engine activity timestamp
Now
All events processed locally
Audit shows minimal metadata only (counts + rule IDs). No message content stored.

Upload handling (StoreBasic vs Pro)

StoreBasic: No file inspection, no OCR. Trust-Prompt only detects the attachment event and shows a warning.
Pro: Adds configurable upload policy controls (allow/warn/block) at the policy layer.

  • StoreBasic: No OCR, no file scanning
  • Pro: Upload policy controls (ALLOW/WARN/BLOCK)

Policies control event behavior, not file scanning.

Upload warning (screenshot)

Real UI capture.

AI prompt data protection – upload warning screenshot (no file inspection, no OCR)

Pro settings controls for AI prompt data protection

Pro adds settings controls designed for stricter environments. Click items below to see what each option means.

  • Pro: Password-protected settings (admin lock)
  • Pro: Upload policy controls
  • Pro: Multi-site scope controls

Settings panel (interactive demo)

Click items for descriptions.

Trust-Prompt · Settings
Pro (licensed)
Password Lock
Protect settings from tampering
Pro
Upload Policy
Allow / Warn / Block for upload events
Pro
Multi-site scope controls
ChatGPT only / All supported / Custom selection
Pro
This is a visual demo. Exact controls may vary by Pro version and release notes.
See Pricing Testing & Quality

AI prompt data protection: StoreBasic vs Pro

Same deterministic engine — Pro adds governance features and multi-site scope (when licensed).

StoreBasic

ChatGPT-only scope · local enforcement

Basic
✓ChatGPT Protection (supported sites)
✓Structured DLP + regional identifiers
✓Redaction (masking)
✓WARN + BLOCK enforcement (race-condition protected)
✕Audit (Pro only)
✕Multi-site scope controls
✕Upload policy controls (ALLOW/WARN/BLOCK)
✕Password/admin lock for settings
Install StoreBasic

Pro

Multi-site scope + governance (licensed)

Pro
✓Everything in StoreBasic
✓Audit (local counters + rule IDs, no prompt text)
✓Multi-site scope controls
✓Upload policy controls (ALLOW/WARN/BLOCK)
✓Password-protected settings (admin lock)
✓License system + activation
View Pro Pricing

Technical validation

Read how we test deterministic behavior, UI enforcement, and race-condition safety.

Testing & Quality How it works